Koha-US KohaCon22, 27 Sep 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Build Cool and Insightful Visualizations with your Koha Data

Description

Presenter:
- Raymond Bryson (Montgomery County Public Libraries (MD) - MCPL)

Slides:
- https://docs.google.com/presentation/d/1HGxUNfl08DVLDmeplbk7d6CCVS0mKUfVayn9rIISayc/edit?usp=sharing

A

Hey everybody I'm Ray Bryson from Montgomery County, public libraries, hi people. um First, just a shout out. uh This is my first cohac con and it's been a wonderful experience so Props to all the people, organizing and doing all the work to make this happen. I appreciate it uh special props to whoever is keeping the schema stuff up to date, love it. Thank you love it.

A

um So I found out this morning that apparently my hotel, my room, was next to a uh a haunted room over at the Eldridge. So I was like oh local history, so I was like oh, the air conditioner all last night. It's super loud, but that was not the case. It was a haunted.

A

um Okay. So let me go into this and actually let me let me set a timer because I'm kind of bad with the timing.

A

All right so yeah we're going to talk about data and visualizations. um Most of the stuff I'm talking about is going to be tool set agnostic. So you can use your your choice of tools, um though I'll I'll also talk a little bit about data outside koha, but primarily the dashboards, and things include, and you know, rely on Co-op stuff. Obviously this is quack all right here, okay.

A

So if you're like me, uh you like to look ahead, so you can use the cheat code, I'm not going to tell you which one it is to get the presentation and go ahead. I think there's about 60-ish slides, um encourage you to cheat. You have my permission to go ahead, so uh some of the goals uh I want you to like leave here with ideas and inspiration to do a few things and not all these will be applicable to everybody, but uh I think there'll at least be some stuff here useful to everybody.

A

um You know we're gonna talk about maybe Identify. Some koha data elements is that up on the screen, let me see oh yeah.

A

Okay, then I'll definitely read it, so um the colors are a little off, but it says- or it's supposed to say, learn what koha data elements might be useful in a branch and or a system-wide dashboard. It's not to be interesting and um we'll talk go a little more.

B

A

um The next slide that you might be able to see uh it's uh provide some ideas and tips on how you can visualize your koha data or other data.

A

um You know using these. You can better understand how your customers interact with your library. I got some good stuff in there for you about that, and then um also you know talk about understanding how a visual, Analytics tool or platform can help staff and Leadership use data to inform decision making. I know this is a increasingly uh important thing.

A

um So we'll talk about that a little bit and then uh you know ideas and tips on capturing data data, preparation and, in my opinion, the thing you can get the most bang for your buck, communicating with Library staff using data. If you, if you can do this, uh your impact, the impact is going to be a lot greater. uh That said, everybody I talked to everywhere, I go. This is like a really really really hard thing to do successfully, all right so introduction.

A

Yes, this one, you can see it's supposed to be like bright, pink and white. So just so, you know I I try to do accessibility right! I may have felt a little a little bit here: okay, so I'm Raymond, Bryson um I come from a web search, analytics background like front-end development type stuff and um now I find myself doing all things. Data and I really really enjoy it.

A

uh It's it's fun and great, uh but really the reason I'm doing this is because I kind of considered the big picture and I'm curious about the customer Journey right so before they get to the catalog or website while they're there after like what's involved. How can we make it better for the customer?

A

How can we get the cool Library stuff that we have get it to them, so they can use it and then like shout their praises and come back more and more that's kind of my in-game there and um I've worked in academic, Federal and public libraries I know: there's a law library back there I've worked in law libraries.

A

um You know: I worked at NIH um I I personally I like public libraries, the best, but all libraries are great and let's go to the next slide. This is. Can you see it? Yes, this sums up my opinion. There's supposed to be little hearts all over the place which they're kind of washed out. So if it's not obvious, I I, heart, libraries, okay, so for uh Montgomery, County Maryland.

A

um This is where I work. There's about 1.1 million residents. We have enrolled in our Public School System about 166 000, K-12 students. There there are more. uh You know a few, a couple thousand ten thousand more in private schools and um I. Think the total like under 18 crowd is about 210 000, so that early literacy stuff we try to focus on too. So we have a lot of people uh we can reach. An interesting thing is 40 of our homes. um There is another language, besides English that is spoken at home.

A

So you know this gives you a hint of about like how diverse our area is, and some things we're working with. uh 59 of our population.

A

um H25 are over has a bachelor's degree or higher and we are just outside Washington DC, and hopefully, yes, okay, so this um is washed out, but we are right outside DC, so DC is essentially right there, which you can't see on this uh screen and then in the yellow up here. This is our County when you zoom in and all those little pink or purple stats, I guess those are our actual libraries.

A

um So this is just for context, because I think context is important, especially if you're going to tell stories with your data. So we have 21 libraries we're open seven days a week as of late February, early March, total items in our collection, we're coming in about 1.9 million and holdings about 1.4, and we have 350 staff members and one one key thing to know about our staff members as I go on. Is we have uh arranged so that everybody has access to our visual Analytics tool which we use Tableau?

A

They all have a view license. So this is part of building a culture of data and increasing data literacy.

A

All right! So can you see this okay? This one shows up the most important slide. Obviously, so uh we left something that I'm not going to mention and came over to kohai and we have our Aspen uh layer on top of that and we came over 10-4 2021 and as far as I can tell it's been the best decision we've made in a long time. Okay, so I think there's context. So our library customers, because I know you know we're a bigger Library.

A

um Just we have as of the last fiscal year. We have about 511 000 registered borrowers and we have about 184 000 active users, and you know for total Circ, it's about 5.2 million um about 2.2 million overdrive and hoopla checkouts, and then, if you throw in everything like in-house use, all other electronic uh resource use, it comes in about 11.6 million. So this is what we're working with as we go forward age distribution.

A

This is a cautionary tell also and making sure your data is comes in okay, um here's the curve here uh got a you know.

A

Almost it's interesting I like this chart by the way, because you can see when people decide to go off to college or leave home, there's that dip and then it climbs back up when everybody comes back and has kids and wants to go to story time uh that big spike cautionary tale when we switched over to a system uh before you didn't need date of birth, and then somebody decided to give everybody the same date of birth. So we do not have an outsized number of 47 year olds, um crap in crap out.

A

So if, if, if uh you can do the whole, you know validation or limit like what can go in What fields. You know you do not need to. If let me put you this way, there are people who are over a thousand years old um they've lived a long time. They love their libraries. That's also very, very wrong that should not be allowed okay and then here is it comes out, okay, but this is a density map, so you just plot all our registered borrowers on here and you can see the hot spots.

A

It's a think of it as similar to a weather map.

A

um We, the closer to the city, the denser, the more users there are, and then our Corridor out of the city, it's kind of a line you see there um this. This is where we have a lot of users.

A

Okay, so what we're building um is basically to sum it up. If you just want to like tune out for five minutes, I can tell you right now build more interactive, less static. We want to move away from cell sheets. We want to move away from PDF documents. We want to move away from static charts things like that. um Not that there's anything wrong with that. I still use all these things.

A

It's just you find that when people can interact and play with the thing, you guys strike me as a bunch of people who like to experiment and play with things. So this would be right up your alley.

A

um We find that people are engaged more with using data to inform their decisions or find holes or places for improvement, so more interactive, less static Okay. This may, let me see if it washed out.

A

Okay, yeah, it's kind of hard to see uh if you, if you go online, if you have a computer, MCL dot link, backslash pres, uh you know the colors might be a little better here. So this is an example of um a system dashboard an overview. So all our staff have access to this.

A

um We also do one. You know we have a system-wide one that Aggregates at the system level, and we have um a branch version of this where people click and choose their Branch, because you know every everybody can get. um uh You know I'm not saying it's a competition, but branches can be competitive and they want their Branch to be the best do the best and they want you to know um so. Brands dashboard everything you see on here um again to promote like a culture of data. Everything here is clickable.

A

You can hover over it. You'll see things um really important for people is it's kind of hard to see up here, but any of the dates you see up there like um in the top left, the checkouts and renewals you're. Seeing the weekly um you've seen this across the weeks, but staff can you know hover over click plus or minus and they could say add I want to see it like monthly. You know, maybe their leader and they're, like I, just need to see it quarterly or fiscal year wise.

A

We, you can do that anybody can do that across everything on here.

A

um So for koha, it's the check out some renewals, the self-checkout percentage and unique active users and we'll go into that a little more later check time. Okay and then we pull in other things too. So programming phone calls. When uh the library was closed because KOA we did um what we called uh holds to go, which is basically like curbside delivery. So we had a sheet. You know like one of the black buttons up there would be uh hold to go.

A

You know it would be there, so people could see because again libraries wanted every. You know they wanted to know how they were doing, but they also wanted everybody else to know how they were doing comparison to each other. So a little friendly competition, so I was going to click through these. uh Try to save some time for questions, uh you know if you click any. So can you see that yeah? So if you were to click any of these black buttons, it would take you there. Here's our checkout page.

A

This is just Co-op um standard stuff, uh also very popular among staff. The checkouts um example why this is kind of cool the date slider, which is on wherever we can put it or wherever I can put it I put it there, because staff like to, for example, do a thing like. Oh we did this new display or we moved our checkout self checkouts in this location. We did it during this time frame.

A

So then they would, you know, move the slider to whatever time frame they're looking at before or after, and then they would try to see. Okay did that have an impact. Did we see a change? Sometimes it's? Yes. Sometimes it's no! Sometimes it's more complicated than that and we'll talk about that. um So that's just checkouts. We have foot traffic uh here, people love to compare again across branches. This is why there's a table there I'm not a fan of the table, but I like compromising if it gets things done.

A

um There's the day of the week is an issue I or people who um are questioning sort of like well. Should we be open this day of the week or that day of the week the day of the week? Kind of breakdowns are very, very, very, very useful and they can encourage discussion or back up certain points or decisions, so just throwing that out there. Okay, we have the active users and we'll go into more detail about this later. It's one of my favorites um and here we're plotting.

A

You know, standard area, charts bar graphs got the slider. We can filter by card type, that sort of thing, uh but we have this uh ZIP code distribution and it's plotted. uh This dark darker area here makes sense if you know the area, um but it's it's useful and we'll go into the more detail about that and here's just programming. This is our second most popular thing. Here we have a heat map, uh the times of programs.

A

This is a very, very, very, very, very, very common question: it's like what time is the best time to do a thing, and so you can look at the Historical information and see um or if you're, trying to manage multiple story times across all 21 branches or Regional areas. You might actually want to look and see where you know are all branches host in story time at 10 o'clock at the same time, across the entire County.

A

Maybe that's a bad idea, maybe you want to you know post them at different hours, different locations, um lots of filtering here and finally, the newer newest edition is our phone logs. This is all mostly automated uh ingestion collection. Everything here is automated, so I'd like to get everything automated or as much as automation as possible after verifying that it's valid and good to go. So you can like, if you were, to glance at this you'd, see a huge Spike and you'd be like what happened in January like that would be my first question.

A

Well, it turns out we decided to um give out code kits and the libraries were basically rushed. I mean calls foot traffic uh checkouts too, to some extent, um so we have a lot of data to back that up and just to know these are these are high level, so you should be able to glance quickly not be like you know, burdened with trying to figure things out.

A

You should be able to glance and understand right, but we for all of these things, I have much more complex in deeper at common projects, workbooks worksheets that allow our staff to really really dig down and deep and and get some insights, but for most people, that's too much. So my really strong recommendation is, if possible, keep it as simple as you can, at least when you first throw this in front of people's eyeballs, otherwise you're going to scare them off all right.

A

So I had mentioned that um one of the things that I think is really important. I mean we just had a session about train the trainers right and like communicating and uh throwing things out there. So communicating and sharing data is super. Super important and we've been like slowly trying to improve this and work on it and no matter what tool set you use these I think some of these tips might be applicable to y'all, and we found that this has been helpful and it works for us.

A

So again can't read this, but what it does say is encourage a culture of data and data literacy. So this is kind of the in-game here um all right, you can see my favorite game so civilization. Six is my favorite game. Anybody out there play that anybody.

A

Yes, my people, okay, I'm, not going to tell you how many hours I've played this, um so everything we're I'm, building uh we're building it to last, not like forever, not like the Great Pyramids right, but in the sense that staff can interact and play with it and not worry about breaking a thing. So every single time I have a discussion, a training or like a presentation or a one-on-one whatever it is, I make it very clear and up front. You cannot break anything.

A

So please explore an experiment and for the people who are more geeky and that's they kind of like breaking things.

A

I taunt them I'm like go ahead, try to break it, bring your A game just come on, let's do it um and every now and then somebody breaks something which is great well for me, it's great um so for most people, just reiterate that you know you can't break it um and at a bare minimum, show how show your staff, how to reset the view whether it's like refresh you know the browser or, if there's a button, to go back to the original view, uh make sure you show that to them um to keep with our gaining stuff here.

A

This is my favorite photo. I ever took um encourage questions. uh This has multiple benefits to the data team, to the culture to the organization, uh but by by encouraging questions and feedback and ideas for improvement, you're you're, essentially starting a conversation right and kind of sharing information.

A

um Really important, like you need to do this across all staff levels like from the director or you know, the board or whatever.

A

It is all the way down to the person who, just this is their first job ever like everybody has valid um opinions, and some people have like great ideas that just pop up and uh I I try to be as transparent and as inclusive as possible across all levels and really what this does is people will ask questions and there's, in my experience, um there's a good chance that you've already answered the question you have the stuff. They just don't necessarily know uh that it's theirs or how to get there.

A

So once you show them where and how to you know to get to it, they become self-sufficient, and then it's sort of like this interesting, like feedback loop, where like they get it and then they tell other people about it and then it just more and more people get sucked up into it and there's ultimately less work for you. But the impact of your data and your insights grow across the organization.

A

I have a little thing about proactive and reactive, but that's just saying like if you have a strategic plan or initiative, and you know that this is on the agenda in the near future, go ahead and start building a thing to address this and make sure your data set can or you have one or you need to expand it or alter it. Make sure your data set can address these particular questions.

A

I, try not to be reactive, but you know when the director needs a thing in a day you know I'm gonna react, okay, so share often and widely uh everyone is all y'all will have different ways of getting out information to your staff. uh We we use news posts which are kind of like blog posts um and an example here that it's kind of washed out on the screen up there. But if you you know you're following along, uh you know we we shared a couple things.

A

um You know like unique activities: map by ZIP code, pretty good Ovid versus covid is like by far the most popular thing. So staff eat that up they, they take these uh images and then they go off to other. You know ala or other conferences and use it and um it's beneficial to them. It's beneficial to us and everybody has a better understanding, um so News Post great I I, really like that idea. It works for us. We have a teams Channel, you guys, might have slack or some other tool.

A

So this is like a real-time communication. um It's focused on data and it's totally open to all staff again that whole idea of being as transparent as inclusive as possible, strongly encourage it. You might find that somebody will pop in and answer the question before you even get around to it, because people get it and they understand it or you may have to redirect the conversation ever so politely in a direction that is beneficial either way. People are talking about the data and using it um a big one. Is you grow?

A

There are tools that will help you do this, but honestly, like at our stage in many many stages for many people, uh a simple like web page would work so we Define and document our data. So we have active users, we've been doing this for a while and we kind of give a definition. So there's no confusion. So we know that staff aren't you know they say. Oh, we have X active users and they don't go off and say yeah for the last three years. No, that's not how it works.

A

We do it yearly, um but we have a definition there for them to reference uh I put in the data sources. The granularity like how often we pull the data in a refresh Target, which is essentially how often they can be sure that the data is updated and refreshed, and this is important if you don't directly connect to a database or if you or, if you can but or if you have multiple data sets outside of koha.

A

You know you're going to need to people are going to need to know when to up when things are like up to date. Okay, oh there was a little. Let me see if I can go back. Okay, well, I'll go back! There's! Basically we make sure there's a question mark on all the dashboards and that question mark people. Can click that and it will take them to the data documentation, page right, so people, you know it's like a help button and it takes them directly to the performance measure.

A

Now this is useful. It's like kind of stops, a lot of random emails or conversations uh before they become a thing um and again it's transparent and everybody can get to it. So let me check the time. Okay, so really um I'm big! On this, like data literacy, data culture thing you could have the best well, curated you could have all your data could be perfect, like no. You know no thousand-year-old zombies!

A

um Good luck! If you have that you could have the perfect data data sets and you could have the perfect visualizations, but if your people don't know how to like use them or why they should use them, it's just a waste of time and resources. So this is kind of what this um is supposed to the story. I'm trying to tell here all right, all right, so unique, active users and there's that bright pink again.

A

um So, just as a reminder here is just from the system dashboard- and this is the thing you could build with a real simple report um pulled from kohai and again our definition here. An active user is basically somebody who's done something with their library card in the last year. I know other places have like a three-year range or two year, or they may actually not include all the category codes or like library card types that we include, but this is our specific definition and has been used since um well.

A

I think it's the next slide. Yes, it's been used since 827 2019.. This is one of the things that we've been able to successfully um Bridge from the ILS that I will not name to koha. The data has lined up and it has worked so what's happening here is we run a report? We capture this information daily and our current data set is about 16.5 million rows.

A

That may be a lot or a little depending on where you're coming from, but uh it seems like a lot, but when you, when you plot it, it's not as much as you think, because, for example, somebody could come every day for a whole year, you know and times 500 000. You suddenly have a lot of rows of data.

A

Okay, so here's this I'm calling it the one simple Quahog report I mean it's really coming from the borrowers table point out here that we use Aspen. So we can use last scene to do things, um but really the most important thing is the last scene, the borrow number and the zip code and you're gonna. What's interesting about this? Is uh we can't really connect to the database here? Because we need to grab the last scene every day because that gets updated every day, so we have to essentially take a snapshot every day.

A

um So I don't know if I mean I'm sure there are other things where you would want to do something similar where you have to take a snapshot every day, so that you can track it over time and I'll go I. Have some other stuff some other slides to talks about this too, but but really run it daily, and then we just kind of count distinct. You know, look for The, Unique, borrow numbers for any given time period, and that gives us, like our total, uh unique active users within that time frame.

A

So if you look at a week, if you have a bar, who does something on Monday Tuesday Wednesday, Thursday, Friday they've been five times in a week that counts as one visit. If you have a bar, who only comes once during that same time frame that's a unique visit, so you just kind of total those up and then once you have that you can start doing things.

A

So here is just a super super simple, like we're: plotting it by ZIP code on a map right um depending on your tool, you you may or may not be able to do this. It may look slightly different, but really the ZIP code is kind of like the highest level of I would call it like Geographic granularity that you might want to focus on, and you know this is cool and all right. But, what's really important is if you start to you're able to tell a story right so and and put this information in context.

A

So let me see if this shows up here. Okay, so what you're looking at here is um that same. You know we're we're mapping unique active users by ZIP code, but we're doing um kind of like a snap, we're looking at it for a different um like date ranges so on the top left. It's pre-coded, um the green stuff is where you know we have a lot of active users again for us. This makes sense within code would hit and we enter the medieval period. It was nothing but horrible things.

A

um You can immediately see it's kind of hard to see here, but you can immediately see there's no green because they're supposed to be green there, but again um there there's no green. It was like radio silence, nothing was happening. Then we entered what in the bottom left quadrant. We enter what uh I'm calling the Enlightenment, and this is where we figured out how to do curbside delivery or holds to go, and now you can start to see our active users pop back up um it's.

A

This was interesting for us to see who was the first to really engage with, holds to go and curbside delivery. um It was not well, I, I won't go into it, but yes, it. This is where it's useful to have on a map to see where people engage in relationship to like uh the story of the context, like the people who engage first are not the people, we thought would necessarily engage first and then the bottom right quadrant is just looking at up to that point, the covid period.

A

um So you could do this too. You may have different like a different story. Maybe it's uh you know before fines within the early stages of fine free uh middle stages, and then you know late stages. You can frame it. However, you want. This has been highly highly highly highly effective in telling a story and getting um uh you know, leadership and you know County the County Council people like that to understand. What's going on so another thing you might want to do is uh blocked users, for example, are your blocked users?

A

Is it Congress is this? Is this? Are your black users in a place where you know low income and poverty is an issue? Well, you know, maybe you want to do something about it and then you can show using some sequence like this: the impact of your changes, so just something to consider here all right. Okay, so let me check here all right, but you can see it's supposed to be blue and red.

A

um So maybe zip code is, you know, not is granular as you want to go. So we have. um We have these things called Equity Focus areas, and these are census, tracts um I've identified 56 of them a bunch of Geeks got together and used like demographic data and plot it and say hey look. These are. These are places within our County where lower income, people of color, who you know, may not necessarily speak English very well right. These are areas we want to Target with Outreach and efforts.

A

um So maybe you want to plot your active user. Other information, on top of this kind of thing, which is a good idea, so the next one we're just showing like our libraries on top of it and I promise. This is all related. um We're PL we're plotting our libraries on top of it. So just alone you can see which libraries may or may not. You know uh want to work on this or focus on these kind of uh issues um like the one in the top right.

A

It's you know up County, it's not in near any Equity Focus area, so you might not want to like have a specific uh your resources dedicated to equity, focused area issues at that particular Library.

A

um But let's say you, let me see if this shows up okay, yeah. Let's say you want to look at um pull your active user and Link. uh You know the people the day people registered for a library card right. You want to kind of look at it that way, so the bottom map is showing for a 10-day period, kind of like where, where people are um it's a heat map, the the red orange, that's the more people, um the green is less. So this is interesting.

A

But then, if you say okay well, let me link like date registered and then compare it to last scene so for like a 10 day period here in the top right you're, seeing a bunch of blue and green dots, and this is by the way, all overlaid on the equity Focus area. You can't see that, because I blew up the dots just because I don't want people to like you know, know our points here. Kind of maintain some privacy um I would never look at it like that.

A

I would make it smaller, but for sharing widely. This is what we're going with. um So the blue dots represent people who registered for a library card within that 10-day range, but have yet to actually come back and do a thing with their library card right and the Green Dot means that they have done something in 2022 and we have it set up that you can pick and choose. You know you can change the when they got their library card.

A

If you look at the whole scope of thing all 511 000, we see some weird stuff, like people who haven't used a library since, like 2 2012, somehow but they're still in our system, which makes no sense whatsoever opportunities for data cleanup, yay, um okay, so this is all cool. This may be useful to tell a story, but there's more right. So here we have our date registered. You know our last scene, but what? If we looked at it by card type, so on the left? It's kind of hard to see on the left.

A

Those are digital cards on the right. Those are regular library cards, so you can see just by looking at the percentages there that um you know, 80 of the people who register for a digital card within that 10-day frame have yet to come back and do a thing in our library, hmm I wonder: what's going on there right, um but compared to regular cards, you know it's it's closer to half and half. So this to me is more reasonable.

A

But again he still begs the question like: what's going on um the thing on the right: uh it's just the absolute numbers, so our digital cards uh a little more than double our regular cards for this time period. Excuse me um again everything here you you all could technically easily do um too.

A

But again, if you put it in context, and you tell a story, this is kind of like what's going on right but I happen to know something I happen to know that um within this time frame uh the county was um basically, they have a program called computer for you, and this program is where we're handing out 50 000 computers to low-income people, people of poverty, people who basically don't have computers at home, which, as you all know like this whole digital divide.

A

This, like you, need the technology in order to access all the library's cool resources. And you know some of these digital resources are not the best on a phone, um so you know again, computers can be highly useful for that and other another reasons. So, what's interesting here is that in order to get a free computer you you have to actually register for a library card. If you don't have one.

A

So if you see that huge Spike of green right, those are people who are either registering a couple days before or the day of uh basically the release of the computers, and they do them in kind of like bundles.

A

um So you have to have a card in order to get it. So this begs a question like: well, you got a card, you got a computer, but we haven't seen you 80 of you disappeared. What's going on so there's room for improvement um communication. Maybe you have a different onboarding process, there's a lot of opportunities to explore there and theoretically, if we kind of look further out and we change our process, we would be able to show oh well when we initially launched.

A

This is what was happening, but then we changed the process and we worked more closely with somebody, and now we can see that, like maybe 60 of people are now coming back to the library and doing things, so it's an example of a useful thing here. um All right. Another interesting thing is um this is show up yes, okay, so here is.

A

This is just basically looking at last seeing data and I wanted to know well back up a second because you have or we have you have millions of rows of data, so you're taking daily snapshots. You will know when somebody comes and then the next time they come. So any in your tool, you can simply create a calculated field um and and say: okay, look! What's the difference between the time they were first seen and the time they were.

A

You know seen the second time or maybe you care about the first time in the 10th time you you can choose these sorts of things, it's up to you right, um but then you kind of map it this way and you can see um for us we're looking at months to repeat visit how long it takes a person to come back uh for a repeat visit to come in through Aspen and kind of interact with our content.

A

um It doesn't tell us the specifics, but again we just really want to know. Were they coming back to us to do a thing um we can change this to look quarterly uh yearly, wouldn't make sense, because we're still kind of newbies there, um quarterly we've got a little but monthly seems to be the sweet spot.

A

uh This is highly useful I. Would you know I the lighter blue? If you look over here, you can actually it looks better. The lighter blue stuff here, if you look at the lighter blue um I, would say like okay, well, look buy if they haven't been back by month three. Well, then we need to get our marketing and Outreach team. They need to step up the game and like reach out and do a thing because we're just going to let these people laugh and they don't.

A

They probably won't come back for who knows how long um okay? So let me check time all right, uh so collecting data questions to ask these are things that uh I kind of generally ask um about everything, but you know I'll just apply to the active user stuff uh one. Does it address a problem or need there's like an infinite amount of data out there? Unfortunately, there's only like one of me or maybe like a handful of people doing this.

A

So in order to prioritize it's good to know like does it, you know, does the stuff the data set or does the data? Will this? Actually, you know solve a problem.

A

um Yes, in our case, because we had no idea who was um I mean we had registered borrowers, but we didn't know like it was actually coming to do a thing in the library. Even at a this high level, we just we had no clue, and now we know- or we have a much better idea um when to collect this comes down to like I, guess the specifics of like the data data models, but really for us active user. It had to be daily because it can stuff can get overwritten daily.

A

So we want to grab it daily. I, don't think we need daily for a lot of things. um You know we can collect. Well, I'll, stop there um and again, another big one was like how much effort to expend is it a manual or automated process? I can tell you right now: we have a Cron job set up. It sends this report to an email and then something sucks up all that stuff and basically adds it to like our data repository and then that gets all like unioned and linked together to build the data set.

A

I can guarantee you. This report runs at, like you know, right after midnight for the previous day, I can guarantee you that I would not get up at midnight to collect active user information I'm, not that dedicated I love libraries, but I'm not getting up at midnight.

A

um So the automation here is crucial, or at least that component of automation is crucial on the flip side, if we ran this, if I had to manually run this report in the morning when I come into work, we would lose um significant data because it turns out surprise, surprise, there's a lot of like night owls and people who interact with the library between the hours of midnight and like 8, 9 A.M, it's actually kind of surprisingly significant or maybe not I, don't know so you know.

A

The other thing is when you want to refresh your visualizations and reports. This largely depends on when you ask staff like what's useful to you and there's that communication thing you gotta be like what's good for y'all, um we can make it happen and then, of course, I have this at the end. Do you plan on linking this to other internal or external data sets I? Would all I would like encourage you all to like consider um like a couple steps ahead?

A

Is everybody becomes more data proficient and data literacy expands like people will start to ask questions like well? Can you link our checkouts to the number of students enrolled in farm programs in our County I'm? Like uh yes, that's possible kinda, you know, but um it's yes, we'll look into it, but these are the kinds of questions that I'm not knocking them. They're kind of they're good questions right, but uh you know it's just you there's a lot more stuff involved, all right, oh tip, so um I inherited some interesting data.

A

I guess is the best way to put it um I, don't think there was like uh there was no will or plan or a state of estate plan or whatever um it just kind of like here. You go enjoy it, um but, but essentially one tip I found is like to establish and use a numerical Library code, because if you have libraries like us, we since I've been there in three years, we've had two libraries change their name like officially change their name. So you know that's kind of a pain.

A

I know in kohai that you, you know Yeah. The code like changing the code is um a thing that can be um a lot of work to clean up, especially if you have a lot of data that use the previous code and that's just in koha. So if you have multiple systems like koha, we have our reservation stuff. We have our program stuff, we have other like Excel sheets, all kinds of things and everybody has a great way to refer to their Library.

A

um It's good to standardize it so numbers it's hard to mess up numbers, um I, encourage establishing a numerical Library code and then finally- and this is maybe more for me- are people who use tools like me, but I really would like encourage you to consider prepending like a year month day, if it's a flat file that you pull down, because when you like link it all together, um you know in your union and you pull it all together.

A

You have this massive 16.5 million row data set and then suddenly you find problems and you're like where the heck did this come from. Well, if you have the file name in there and it has a year month day, it can help you investigate and kind of troubleshoot, so I found. This has saved me so much time um do I have time. Yes, all right, so other cool things if time permits all right.

A

So, um okay, does it show up here, yeah, kinda, okay, I'll read it so here we're just looking at um staff and self checkout Transit transactions at these three libraries on a specific day. So one of the things is, you know we we started, uh we opened up on Sundays um and we want to Monitor and understand, like you know, what's going on with Sundays? Are people coming to use a library on Sundays like intuitively some people be like yeah? It's the weekend.

A

People work during the week, of course, they're going to come and then, on the other hand, other people who work in the library like we don't see anybody come in on the weekends. So you know it's just here's where data can kind of help clarify some things, um so this particular chart is looking at. um You know the trends so that little the dotted line is the trend line um for each Library. These three libraries on a Sunday since we started Sunday hours.

A

So here it's saying for staff and self checkouts that you know the trend is it's increasing, maybe not at a super like exponential rate, but it is increasing and it makes sense customers. You know they are using the library on Sundays uh the actual. um This is like an export, but the actual dashboard or project itself. uh You can filter by day of week, Branch interface.

A

You can come combine the day of the week like let's say you wanted to compare weekends to weekdays or if you had like, maybe you have weird hours on Monday Wednesday Friday across your system uh on Tuesday and Thursdays. You could compare hours that way. So these are things you can do uh and I find them or we find them very useful.

A

um So maybe day of week is a little too the bucket's too big I, guess maybe you care about uh transactions per hour open. So uh you could do this.

A

uh The the caveat here is that you need to have some way of knowing your your hours open right for each branch or if it's a system you might maybe you could estimate if your system has the same hours, but we do track hours open, like our our managers will put in hours open, and this is useful because sometimes you know it snows, where I'm from so we close down or it gets super hot and the air conditioner doesn't work we close down.

A

um So we do that. Anyway, you can filter transactions by day of week. Branch or interface and what's cool here is um you know you can see that here staff transactions, so your staff they're, actually checking stuff out on their on Saturdays. This is their busiest day right and then you know in the week Tuesdays Thursdays, maybe Friday, um those are their other busy days, but you get a completely different story.

A

When you look at um customers who just bypass staff and say you know what you know, I can use the self-checkout machine in fact, I'm going to use it because I got a kid who's ready to go home and I'm stressed out. um So here's a different picture. Does it show up yeah, so here you're, seeing Saturday is still by far the most popular day um across our system, but then um Sunday's clocking is number two followed by Friday, so you get a completely different um perspective.

A

So you know this might be important for Staffing things like that. uh Then you can combine them. Let me try to combine them there. We go so you can combine them so then you can see it in context um again. Saturdays are rocking it followed by Sundays and then Fridays. So this is useful. um We can change the date date range if we want to. You can look at a specific week or not it's up to you um and you can look at your specific branches if you want to or combinations of branches.

A

If you wanted to look at like Regional areas, um this is all useful stuff and then finally, we have self-checkout percentage as a system. uh This is something we're we're trying to encourage self-checkout um and you can see kind of when we migrated to koha that you know staff was a lot more involved. You know it was all new to people, staff and customers. So there's a lot more staff interaction, but as people became more accustomed to it and staff became better at communicating uh about koha.

A

We see our cell check out percentage rise and um this is useful for us, which is why it's on a branch dashboard and just to go over my current toolbox and again this stuff should be to everything. I talked about is mostly tool set. Agnostic like you, can choose your tool. I know. Somebody had asked a question earlier in the week about power.

A

Bi, like you, could probably do all this in power bi, if you, if you want to pay for the licenses things like that, um so here's just the things that I'm using you know: Tableau Tableau, prep, arcgis policy map, courtesy of our state estate, Grant, uh I, good, all Excel, to quickly view things, um uh Google, Sheets, uh obviously, KOA reports, kind of jobs. We use ESP for some things, we're we're in a unique position, our relationship with our I.T Department, um but they do have an ESB.

A

We could use to link and connect data and pull it out and store it and drop it in interesting places. uh Then we use Microsoft teams in SharePoint, because our county is a Microsoft kind of County, For, Better or Worse, and this is the slide that actually works. Okay, that is all folks.

A

All right, it looks like I, um okay, I just want to say, I definitely went under my time like, which is good, so.

A

Okay, oh I, see a chat well, I'll hold it.

C

All right we're going to answer some questions here. um I just have to zoom. So if George, can you pull up the or Jason whenever you pull up the YouTube and see if there are any questions on there? uh Let me get to the right place here.

C

So someone was asking.

C

Where'd you go not easy to navigate with a mic in my hand,.

C

It's a little better.

C

Someone was asking: can we integrate one search with koha and Google Voice Assistant to koha OPAC.

A

I have no idea. I will pass that to somebody who's aware of that maybe buy water people I have I, don't know.

C

Can we integrate one search with koha and Google Voice Assistant to koha OPEC.

C

C

Okay, yeah, you might need to uh write this and clarify on that question. They also asked which is best open source web scale, discovery that suitable that is suitable for koha.

A

I'm gonna pass that.

C

Okay, um do you factor in the zip plus four or truncate it to only the zip.

A

um No well we're just using five digits.

A

Yeah yeah we have, we like, even even in the past, we've pretty much locked off anything past five. So for our purposes, that's that works for zip code. um If we needed- let's say you know higher Precision. We have other ways besides zip code to to get that information.

C

And then there was a comment uh this uh year month. Date is the best in that order, uh lots of applauses. um Someone said that was great. Thank you and.

C

More applauses and thanks team any comments on YouTube or questions.

D

um There's a comment here from Dwight.

D

So his first question is: is the digital card an app it's just you're, just using the it's just the borrower category in koha, that's used for the self-registration right or probably yeah, it's just a type of library card they use at his library. And then um the next question is the digital card just for erases, or is it for everything.

E

D

So if they have a digital library card, they can place items on request and then, when they come to the library, they need to change their card up from a digital resources card to uh to a regular library card.

D

And Dwight also says that the low income area map was very interesting. That was part of policy map right.

D

Yeah here hold on.

A

So that um there may be a link, I provided I can get the link. But essentially what happened was our as part of our County. uh We have like some planning people and they went and grabbed a whole bunch of census and ACS stuff.

A

um They pulled pulled all that stuff together and then they kind of had their own sort of formula I guess you could say, and they ranked um certain uh elements and then they said okay of all this, because they ranked all the census tracks right.

A

um They set up those the ones that scored within this range were considering these Equity Focus areas. So maybe policymap has something like that, which would have probably saved them a lot of time.

A

But uh that's how well that's where I got the information from our our County Planning people.

D

The policy map was kind of a question for me not from Dwight, because I use policy map and for some of our libraries and it does have similar data. But it sounds like what you guys are using is really kind of bespoke for your library for your area um and then there's one other question here: sfpl I, don't know who that is, but they say if libraries do not currently do much with data visual visualization. Where would you recommend that they start.

A

uh Maybe, like many people in here I'm a big fan of uh just picking a thing, a tool and experimenting just to go with it, uh but if we um so in Maryland this all kind of started, because we received a grant for Tableau right, so we sort of had our tools set up and we we had a community uh of people who had been using Tableau for at least a couple years.

A

um So if you can find a community in your region, State even County, that's where I would start because, especially for libraries, because libraries have very you know, kind of unique questions that maybe the private sector might not care about or have um I would start there um and then, of course, I would pick a tool that has a really good a really good community and documentation and start experimenting with some data sets like just go to town right. Ask your leadership or whoever and say: what's what do we?

A

What questions do we want to answer the most and pick one or two and then just see what you can do.

C

E

With your recent edition of Sundays and your hourly chart, you had there: where did where do you see the checkouts being most prevalent on Sundays? Is it the last hour? Is it the first hour? How do you see that mapping out for you.

A

I have done this and I don't want to say the wrong thing. um I will say that I feel like the the lunchtime hours or like kind of where it dips there's like a mad rush, and then a spike before it closes is I I. This is what I recall um I've done like hundreds of these things and they all kind of blur together uh but yeah, but that's a good idea for you know we were looking daily right, but it's totally maybe sensible, also to do like sort of like an hourly type thing.

A

um The way we did with like phone calls, for example, or the programming where you know each hour, like you, use a heat map. You know the redder. It is that's the more stuff. That's happening totally possible to do with Circ, because um uh basically, the date, the time date stamp. um We could do that. So maybe I should thanks.

F

Quick question, so uh you had mentioned that there was programming data. uh Do you know how the individual libraries report that to you? Is it just like an Excel sheet that has that information or do they like? Has it in koha somewhere or something along those lines.

B

A

Programming I feel, like every library has this problem, so we use uh communico and uh before that we used events right, um but with communico we well here's. What really happens? The libraries will um create the program. We have certain requirements for audience tags like primary audience tags. You know, event type all that kind of stuff right and then what's supposed to happen is immediately after the event they drop in the attendance, um and we again we document we have documentation and saying this is here's some.

A

This is what you're supposed to do and for the most part, that's what happens, but you know with new people. Sometimes we get bad data and you know we kind of have to remind people often, but they put this in and then I just at the end of the month. This is where that refresh cycle comes into play.

A

um What happens? Is the managers or somebody a supervisor, is supposed to verify that all the stuff's in there um and then after I get you know, green light, I, say: okay, I'm, going to pull the entire month's worth of stuff down and I, pull like literally everything down like I. Think um it's like the a a so it goes through the alphabet once it doesn't quite go through twice so I think there may be, like you know, in the 40s, so 40 some odd Fields.

A

So we pull all that down and then from there I I clean up and do all kinds of magical stuff, um and you know report back out. So that's kind of our process. I do know that other libraries use Google, Sheets um and that's been successful and it may be even easier, especially since you're using I think metabase right. So you might be able to connect to Sheets a lot easier um we may be able to.

A

If you use communico I know, people have used their API to drop that information into say orange boy or some other place, uh but for us it's just. The manual pool is kind of the easiest thing, especially considering the amount of um questionable data that I keep seeing.

D

um I wanted to ask about. You mentioned data literacy. um What kinds of things do you teach your stuff about? You know how to interpret the data and how to make sense of it, because you know one of the things that I can easily that I've experienced one of the things that I've experienced. A lot is like libraries, you mentioned, being open on Sunday and I've, experienced libraries that have had a change in hours and then, like three weeks later, they look at the data and they say well nobody's coming on Sunday.

D

We need to just you know the hell with that. We need to close on Sunday again and I always tell them. You know it's going to take at least two years before we have any good data, so I'm wondering how you address those kinds of issues with your data literacy training there yeah.

A

um Are you sure you don't work in Montgomery, County, Public, Libraries, I've heard this question before it's come up multiple times we have like an AMA. uh You know ask me anything uh with leadership in this question. This exact question has come up multiple times since we opened on Sundays.

A

um So one one thing uh that well: okay, two things: uh data literacy, I kind of there's like the technical component right. So do you know how to work whatever tool you're using like that's just like okay, uh but then there's the bigger picture of like? Can you look at the data or look at the data in the context of a question right so for the the foot traffic question or or the Sunday hours question they give you an example. uh Some people would look at foot traffic and only foot traffic.

A

So here we have to say: well you know that tells you a story, but you should also maybe consider uh transactions um where they're programming, um like the Staffing on that particular day right. It may feel like nobody's there, but you may have more people around or vice versa, um so one one step is just to like get people into really looking at the full picture and not just cherry picking the data that proves what they're looking at like.

A

Ideally, the data is there for you to disprove the thing that you're trying to say like it sounds weird, but really I want the data to disprove what I my my theory or my intuition about this, so I encourage people like use it to sort of not prove what you're saying but to disprove, find ways to disprove what you're saying so.

C

I think too I'll add to that is just that. You know when we changed ours, it's important to know whether or not did you do a soft opening or or did you get the word out or something like that, so foreign.

D

I was just mostly wondering about that part of train about part of data literacy about what to do with the data now that you've got it rather than rather than how do you address those issues? But you know how do you address training people on how to interpret the data that was really more? What I was going at, and the Sunday thing was just an example that I've experienced I've experienced a lot.

A

uh One thing that we are getting ready to do and prioritize is like this concept of onboarding um up until recently.

A

You know most of this didn't exist in our library um and it might not exist in most libraries, but we're going to start working on building an onboarding process to work on that, like literacy, data literacy and also like the technical side, like you know, here's the basics of what you need to do sort of more formalize that and get it up in people's faces because really I feel like staff, are doing more with less time kind of so I don't want to overburden them, but I definitely want to give them.

A

um You know a way to to get or ramp up their understanding of data and use of it. So I don't know exactly what that onboarding process will look like. But um if anybody has ideas, I'm I'm more than happy to hear them.

C

Okay, one more question.

G

So mine is kind of similar to George's, but it's more. um How do you use that data to kind of look at maybe those hours that are being or and days that are being seen more as well as like when people are coming to programming more and help um use that to decide when to maybe change hours or to decide hey? Maybe we should do programming on this day instead of this day? How do you kind of Express that and find that out.

A

uh To to start answering that question, the most important thing is to have buy-in with your leadership like, if they're, the ones that are going to make the decision uh they're good yeah. If you're gonna make the decision, then it's good to have somebody Express a need to use data to understand the problem and maybe like kind of uh explore it.

A

um So I didn't really talk about buying from leadership, but one reason we're increasingly successful here is because our leadership is like. Yes, we can use data to do things and it's showing us and really um that okay I'm not going to go there and show you the picture but but really like. If you can get the data and visualize it in a way.

A

That's like digestible to people who are essentially, this is not their thing if they can glance at it and kind of understand, you're going to be like light years ahead of just saying in a sentence or you know a table like look, here's our Circ for 2022, and this is what it was last year. If you can show a Baseline and then progress, um then that's kind of where I would start.

A

um I didn't talk about bass lines, I'm a big fan of bass lines like establish a baseline, make a change and observe how that affected. The Baseline.

C

All right Raymond. Thank you very much for your presentation. Oh okay,.

H

Yeah I'm sorry um I was just interested if uh you ever thought about sharing that data like in an open data Repository, because what really strikes me by a surprise since I've since I come here, is how uh how you use your data, really to make business decisions and is completely Unthinkable in Germany, because of our gdpr and privacy laws, and so on and so forth and I would be really interested.

H

If we, maybe in the community or core us, could make a coordinated effort to bring this data to open data repositories, because I also see an opportunity for machine learning and with the data sets that we have in Germany. For example. This is just impossible.

E

A

I, like the machine learning idea, I have I want to do a lot of text analysis of stuff, but anyway, um to answer your question for our specific um County we do have well. We we have to submit certain uh data sets and and document them into sort of like a county-wide repository which, which they're open access anybody can get them do we include everything that you saw up there? No, no! No, because we want to respect. You know pii things like that, but we we can Aggregate and then push it out.

A

So a lot, almost everything you saw up there. This is dependent on like row level like row level data like many many millions of rows of data right we're not going to give that to the county like that, it's just probably too much. They don't need it, um but uh for the wider thing you know we have like imls PLS, like they sort of do these.

A

um They collect nationally certain data elements and that's also something I guess I got to do. um uh Your mileage may vary because again everybody defines things differently. um One of my beefs like registered bars. What's that mean to you?

A

It may mean something completely different to somebody else, um but you know if you want, if you wanted uh some data to play with check out the item list, I like pulled down all their data for the last like I, don't know, 10 years, eight years and I dropped it in a dashboard and then I like filtered it filtered out everybody else except our County, so that we could have our own historical stuff. But then you know it's there to play with is what I'm saying: okay.

B

Well, if he gets to ask a question: I do I live in Montgomery County in Silver Spring before got married and the house became too small. I actually lived in one of the equity Focus zones in 20910 yeah.

B

ah But uh first of all, you do a magnificent job in the in the county, which makes the astonishingly High state.

B

And County Income Tax I pay worthwhile taxes, pay for civilization, but um I think it would be a good use of my and other tax dollars for uh the Montgomery County Public Library System, to consider putting in a bid for next year's uh KOA us conference.

A

As a fellow Montgomery, County resident I agree.

C

Okay, I think that's the end of our questions here and again. Thank you, Raymond for everything, great presentation.