 Thanks everybody for being here. I appreciate it. As Justin mentioned, my name is Brad Stevenson. I'm director of marketing and I just wanted to briefly welcome everyone and talk a little bit about how this project was conceptualized and the connection to the exhibition. So I started as director of marketing with Carnegie Museum of Art in November. Very shortly after I started we began having meetings about our Hillman photography initiative which explores visual culture and it's coming toward the close of its first year of activity. The exhibition Distant Feel by artist Antoine Katala is one of the initiatives that came out of HPI. As we were talking about ways to promote Distant Feel which is an examination of how people respond and react to imagery online today, we were thinking about ways that our marketing could also allow people to respond and react to imagery online and what we might be able to do with that. So as I started thinking about this idea of responding emotionally in a distant way, I started thinking about the golf tower because I'm always fascinated by the weather beacon that's on top of the golf tower. How many of you know that there's a weather beacon on top of the golf tower? So KDKA runs a weather beacon with like six layers. It's this code that nobody can ever really understand what the weather is based on looking at the thing. So but it is a fascinating thing that they do. It used to be blue means it's going to rain and red means it's going to be hot or something like that. It was very very simple and very very basic and now it's the six tiered barometric pressure thing that you have to be some sort of a space-aged pirate to really understand what the hell they're talking about. But I love the tower. I can see the tower from my living room window and I'm always fascinated by the light up and also sports teams when they score it will light up for every goal that the Penguin score for instance or when the Steelers score. So we called the we called Larry Walsh who is the COO of the golf tower of rugby Realty which owns the golf tower and we asked him about it. We said is this possible and he said yes absolutely every bulb on the golf tower is basically controllable through a VPN. And so we were like this is this is amazing. This is fantastic. It's exactly what we want to do. The initial concept was that we would do something where we push out imagery from the museum and have people react and based on their reactions we would light up certain colors on the tower. Well David was already here working with us on an art tracks project which is about provenance which is about the history of an artwork and linked open data and it's a really fascinating project that he should talk about again to you guys at some point because it's really amazing but we won't get into the weeds on that today but he was already here working with the museum. So previously I had asked him hey do you want to help redevelop our blog and he said no thank you boring. So I went back to him with this and I said okay how about this how about we light up the golf tower using you know people's sentiment here in Pittsburgh and he said that sounds great that's more along the lines of what I'd be interested in. And so but he said instead of doing something where we push things out and try to get people to participate in something that we create why don't we just use data that's already out there. It's always easier to take advantage of things that are already out there than it is to try to create something that people have to participate in and I said that's fantastic let's go. So that's how the idea of using comments on Instagram in the Pittsburgh region and measuring the sentiment of those comments and lighting up the golf tower positive or negative came to be. It was during a maybe a 30 minute conversation that David and I had and we had maybe three weeks to put this whole thing together and without David there's absolutely no way that this would have ever happened in the way that it did. So what I want to say about this is that it's important for us as it's important for me as a marketing director that the promotional things that we do I love hype I love these kind of big things that are super fun and that have a tech aspect to them. I also like to connect them to the art and it was important to me that this project connected to the art and it absolutely does. So I would invite you after this to spend a little time in forum gallery you can get into that gallery for free we're open till five take a look at distant feel check it out if any of you want to come back to the museum and spend a little time you can tweet to the museum or send us a Facebook message and I'll get you a free pass for whatever day you'd like to come back I'm happy to do that for you so you can spend because we do close at five and I want you guys to have an opportunity to spend some time here if you you know especially if you haven't been here before. So that being said I'd like to introduce David Newberry some of you know him in real life some of you know him as worker gnome on Twitter and again without him we wouldn't have been able to do this so he'll get into the nitty gritty of how this actually happened. So here you go David and we're good. Hi there everybody I'm David Newberry thank you all for coming out and thank you very much Brad for the introduction it's really nice to be here and the little bit about me I'm David Newberry I'm a developer here in Pittsburgh and my background is sort of in this combination of film production and computer science and hardware software interactivity. I worked for a while before I came to Pittsburgh I was a I'd been working as a computer developer I'd also been working as an animator out in Chicago. I did a bunch of work for PBS out there a bunch of the ad agencies out there and I've also always sort of done computer science work as well and as film became more and more linked to computers the work I was doing became more and more sort of about this integration between technology and the integration that and storytelling and sort of the way you can use hardware and you can use software and you can use technology to sort of take these stories that exist and show them in new ways not necessarily animation per se but if you can take animation and you can move it into the physical world or you can take computer science you can move it into the physical world and you can combine all of these things together. So I've been doing that sort of work here in Pittsburgh for a while I work with Ion Tank in East Liberty I've done some work with Deep Local down in the Strip District I've been doing some work with Carnegie Mellon and as Brad said I've been doing a lot of work here at the museum both on art tracks which is this large sort of research project involving taking the history of artworks also this project here and as he was saying when we worked on this project you know it was this combination of the Carnegie Museum and their idea and their marketing campaign and sort of Brad's idea that we should do something bigger than the normal marketing we did and the Larra Russian Run would be real who were the people who had the tower and were really really generous to let us get access to this and help us so much in making this work and me who and my little design work at home studios so the golf tower we all know it a little bit about its history it was the headquarters for the golf oil company which was based here in Pittsburgh when it was built it with the tallest building in Pittsburgh it was also the tallest building in Pennsylvania from 1932 when it was built until about 1970 this huge 44-story building and one of the things that have always been a big feature of this building is its lights originally when they built the building they built it with neon wrapping all around the parapets of the building red and blue neon and if it was red it was going to be good weather for that day if it was blue it was going to rain or it was going to snow and when the lights were you know steady it was getting warmer and nicer and when they started flashing it was getting colder and they started this with when they put the building up in the 30s and they kept it going until about the 1970s and in the 1970s what happened is we had the oil crisis here in in the States and it turns out that if you run a giant oil company and there's a huge energy crisis the best thing to do is not have massive all the time on lights all around your building it's sort of bad public relations and so they got rid of all of the neon and they just took that little tiny window up at the top and they put a big sodium vapor light up there and would flash out the colors to keep it going one of the other really fun stories I learned when doing research for this project was you know how Brad was saying for all sports events the lights flash and you know celebrate Pittsburgh sports victories it turns out what happened was the afternoon and night receptionist at the golf tower was a huge Pirates fan and so she would sit there at the reception desk and listen to the baseball games and every time anything exciting happened she just had a switch on her desk and would toggle the switch because she was excited so for a while there was this the the radio announcers would actually every time that happened they'd hit a home run they'd go flash the beam Regina this one's out of here and she became this little minor celebrity in the 90s about this flashing the lights so in 2012 the building changed hands and they decided that they wanted to bring back the full lights for the tower and there was this large project to stick LED lighting in particular color kinetics fixtures all around the building and so rimming it they have these now individually controllable 185 light fixtures that are addressable at one foot increments so every foot on each of the six levels is an individually controllable rgb sort of pixel if you know anything about dmx dmx is a lighting control software there's actually five universes of lights up there a universe is 512 channels so there's about 2 000 channels of individually controllable lights up there and they end up using a whole lot I mean this they're super power efficient they're about a quarter of the power of the original sodium vapor lights and this is if you ever wanted to know how to tell the weather this is the chart of what's going on there I still don't understand it so as Brad was saying our project was how do you show this sort of feeling at a distance because of the Anton Cutella show this was the idea we had and the conversations that Brad and I had about how to do this and so this was our original thing the original idea that you know Brad came to me with and said how can we do this it was let's find tweets that have images because this is a visual show it's all about images and they're connected to emotion we want to make sure that we're talking about images they wanted to find what emotion that had because the other thing we're talking about in this show is the emotion of these things and we wanted to use that to sort of figure out the mood of Pittsburgh what people were thinking what people were feeling and showed on the golf tower and so we sort of had this big thing and we're not the first people to do things similar to this um in 2012 there was a they did this at the Olympics using the big um Ferris wheel and the London Eye um but no one's ever really done it with this image hook um basically it's been done with sort of tweets before um but we were trying to figure out how to incorporate imagery into this so we started out looking at Twitter and the problem with Twitter is that it's really hard to sort of figure out imagery on Twitter there's not a really good system because images aren't native to Twitter they don't it's not easy to search for those things and so we switched to Instagram um for a couple reasons one is because if you're talking about images there's no better place to do it than Instagram the other thing is that Instagram has this really really nice API that lets you sort of set up push notifications for locations where things happen it's a sort of a pub sub web you know web hook thing you say I'm interested in this location area and it will give you notifications for it and doing my research I was like this is great this is exactly what we need to make this work and so we then had this discussion which is how do we choose which ones to show um and we talked about messages to the museum uh we talked about maybe we should do hashtags uh maybe we should search for particular words across Pittsburgh for the museum and the problem we had with all of these is um what if nobody participates because this is a new idea we haven't actually tried this before and the most embarrassing thing to do would be to put all of this work into a big project and say and then nobody engages with it and we're talking about updating the golf tower every five minutes for three days that was sort of what the golf tower said we could do um that's a lot of interaction that has to happen and so we decided that using the sort of location based thing is great because if it doesn't you know they have this really nice api we can do it and it turns out that all we have to do is exist in Pittsburgh you know you don't if you can give people a way to check your work to make sure whether or not you're what you're doing is working um you give people an you give yourself an opportunity to fail publicly um no one else is going to scrape all of the tweets you know all the instagram photos in Pittsburgh compare their emotions see whether or not it's actually working it's enough that they can get the general idea and the story is really really good here so uh i don't know if you've ever done anything on the internet but it turns out that the internet is not always as good as the internet should be um there's this really nice real-time update thing and you know you get these notifications and they've got this really nice document that explains exactly how to use it and you start using it and all of a sudden you realize that it doesn't actually do anything i set it up i got it working and it turns out that it only sends you results about every three hours about three results which means that it doesn't it's totally useless because i can't we can't use this but it turns out that one of the really nice things that i've learned from the past six or seven years working on projects like this is you don't actually have to do things the right way to make it work you're allowed to fake things you're allowed to sort of say okay i know what result i want it doesn't have to work the way that you should do it if you wanted to do it right so i did a little math and figured out that the instagram api limit gives me 5 000 requests an hour and if you do some math that means that i can if i'm querying for every update every 30 seconds i can get to about 40 locations you can also query for things that are happening within a five kilometer radius and if you take alligating county and you do some math on that turns out there's about 24 circles in alligating county um if you take circles there if a five kilometer radius which means we can totally do this just by brute forces search every circle in pittsburgh and aggregate the results um it's sort of okay to do things the wrong way for project like this because we don't have to scale you know a lot of you work on internet things you're worried about what happens if you get more traffic um they're not going to build another golf tower i'm not going to have to do this twice in the same place as long as it works sort of it works so this is where we pull other locations from um i don't know if you can see it out there that's a map of pittsburgh those are the gps coordinates and we would just every 30 seconds we would go through this entire list of points and collect all of the photos that happen in that area it's not actually quite alligating county but it's pretty close and it works we get data um for every photo that you get on instagram you get this huge chunk of jason with all sorts of bits of data throughout it and we were getting hundreds of these an hour um people take a lot of photos and instagram is really really good at aggregating so what we decided to do our first idea um and one of these ideas was we now have this photo data wouldn't it be really really cool if we could take the photos and show the photos in the lights of the golf tower you know take the average colors aggregated across use it as a low pixel display show it um and we started talking to larry and the people at the golf tower and as we started talking we realized that it's totally possible to do this they've got a really really nice system up there but that coordination was going to start taking time we had to go over their vpn we'd have to build new hooks into their system to do these sort of dynamic color changing things um it takes a lot of compute power to sort of analyze hundreds of photos and display them and but really the problem was this storytelling problem um i think it's really really cool to use a building as a low pixel display but the whole point of here is we had to sort of tell this story about emotion and photography and pizzerberg and if you have to go and sort of give a technical talk about low pixel displays to every single person who's engaging with this you're it's not going to work you it's not the story that you can tell you know it's only really clever if people get what you're doing and you know wired would think this is really cool if you could sit down and give them a long interview about it the trip doesn't really care um you only get about five seconds of someone's attention for projects like this and if they can't figure it out in five seconds you've completely lost it so this was our second idea it was the let's take figure out the mood of the city let's put it up there and display it um and so we started thinking about well this is an easier story to tell but what data do we have um and it's you can't actually at least i don't know anyone who can take a photo and extract an emotion from the content of a photo that's probably three to five years out image recognition software we weren't going to reinvent computer vision to solve this project in three weeks so we were going to use the text the comments the captions all the stuff from instagram that was textual um and what we're going to use is a software tool called sentiment analysis which takes it the text runs it through a bunch of algorithms and there's a program called the stanford natural language processing program which is incredibly powerful um best of breed does all sorts of amazing analysis you know you can do sentiment analysis you can do tokening all parts of speech uh automatic question answering if you're really interested in extracting all the biomedical information out of a book you can just they have a library for doing that um but it's also this huge monolithic java application that is not particularly fast it's incredibly powerful and we're dealing with hundreds and hundreds of things a second that i have to figure out where it's going to run somewhere so it turns out that there's also these sort of much smaller ruby gems that are out there that basically take a block of text and run it across lists of words and do very basic analysis on this um for every word that you get every sentence you get they sort of look through all the words that exist in there someone's gone through and tagged a bunch of words with emotional scores and what you get back is a positive or negative integer um that talks about what it thinks it is and they're none of them are really good but if you take a couple of them you average the scores you get together you end up with something that begins to work well enough you can sort of say i'm drinking the greatest milkshake ever and it comes out as a plus one that's a good score you can say oh the world's a sad hard lonely place and that comes out at about a negative point three um you can say poof loves helping us change the sheets and that comes out of about a point eight uh you can also say that the battery in my f150 resembles a zebra and that's about a negative point three um if it works well enough it's good enough you don't we're not launching rockets if we end up miscategorizing something nobody cares you just have to let it be what it's going to be so we had this thing but we still have sort of a story problem here which is if you know what's happening this works really really well but in order to know that blue means that the city is feeling very hungry right now you need to have a legend you need to go out and look it up every time it's the same problem they have with the weather beacon which is no one knows how to read it everyone thinks it's really cool that it tells the weather but no one actually knows what the weather is so this is what we ended up with which was let's let's think of what the simplest thing is that we have here we have positive emotion and we have negative emotion and if we use green for positive and red for negative everybody recognizes those feelings and if we treat the sides of the building as bar graphs you can say okay we've got six green and two red pittsburgh is happy we've got four red and two green pittsburgh is upset and it's a really easy story to tell the technical implementation is not particularly complicated we don't you know we have i think 36 states that can be possibly in we can hard code those but we can tell the story and anyone you go to you say hey as golf tower gets greener pittsburgh is happier and you don't need to look anything up so technology and for those of you who are here for the art you've got about 90 seconds to check your email or do something other than this because this is the sort of nerdy part of it we built the software it's all ruby-based we've got a Sinatra back end serving up a lot of the content got a redis data store in there to sort of hold the content as we aggregate it and everything is hosted on heroku heroku is a big cloud platform most of you probably know about it we were running two workers on heroku one of them was a ruby script that did that actual brute force go and analyze all of the tweets as they came in you know query the instagram api get the results back built her down things that had comments do sentiment analysis on it calculate average scores for that and dump it into redis um you know it would then it would do this 20 times it would then just sleep for the next 30 seconds and come back again this doesn't scale it's not web scale but we're not trying to get there we were storing the two most relevant instagrams the most positive and the least positive for every 30 second period the last 10 aggregate stores between them and additional data for metrics and post analysis the other job we had was the web server running Sinatra running web page also running an api to do sort of real-time updates of what these scores were so that the web page could be dynamic and that's the really quick technical implementation i'm happy to nerd out about that anyone after this but there's all sorts of other interesting things um not everybody lives downtown not everyone can look at their living room window and see the golf tower out there so we built a web page so that you could see what was happening and also to sort of explain this and on that we had these three representations of the data that we were getting three different ways to look at the same data one of them was this graph that showed the average score over everything at 30 second increments over time so you could watch the mood of Pittsburgh go up and go down and you and it auto refreshed so people could see it changing and know that this is happening in real time in this dynamic we had the graphic of the tower which was showing what the tower's positions were at any point so that people could know what they were supposed to be looking for and if you couldn't see it you could at least know that it was happening and then we were showing the most positive and the least positive instagram photo of the past 30 seconds so that people could see sort of where these scores were coming from and this was our storytelling part this is how we were explaining what we were doing and so we started mocking this thing up and testing it and making sure that it would actually work and the first thing that we had the problem we had is I've done other projects like this we're doing content curation and sort of taking public data from the internet and putting it you know into an exhibit and we have the problem that the internet is the internet and it turns out that one of the most positive people on the internet who posts Instagram posts is cheerleaders strip club and it immediately started coming up as like hey everything's wonderful you should come down and see us and I freaked out because all of a sudden I'm you know I'm now posting this on the museum's website and I went to Brad and was like Brad I'm a little nervous about this and Brad reminded me that we work at a museum and there's all sorts of interesting things that we hang on the wall when we talk about art and the museum is really really comfortable with the idea that that content that people create is content and that it's not our job to sort of choose what content to show which makes my life way easier because yeah it made me a whole lot less nervous another nerd thing so we start getting this data and we're getting all this information and we're trying to figure out what scores we're going to show and I started out you know I was like we're going to do this right let's do a rolling two hour average of the average scores Pittsburgh so we can give really interesting reliable data about the mood of Pittsburgh and it turns out that Pittsburgh is really really consistently positive people don't post photos when they're really unhappy you don't say oh my dog just bit me I'm going to put that on Instagram your Instagram photos tend to be things that you're excited about oh look at the lovely sunset oh I have a birthday party and I got a cake um which meant that these scores stayed really really consistently about 60 positive 20 negative about 20 neutral and that makes really really boring data it's really accurate data but the whole point of this isn't to be accurate the point of this is to show interesting things so it turns out that if you use sort of a three minute rolling average there's enough noise in the system that a three minute rolling average shows you roughly things are positive but you get these spikes these changes and it tells the story better it's not quite as accurate but again we don't really care about that and so we now we're getting our data from Instagram we're doing analysis on it we're playing on the web the last thing that we need to do is actually get it up to the tower and so to do that we use the museum's high tech data center that we have in the basement that stores all of our networking equipment and we ran it down on that and by high tech storage system in the basement I mean we had a mac and money running in the data lab and I had the the terrifying fear that our intern would come in and start editing video while this was running and turn off the golf tower so I put post-it notes all over everything saying don't touch this don't touch this um and the other story here and the other thing you think about when you're building this is it turns out that it's really hard to test this kind of software you can do what you want but you don't get like a three week burning period of trying the golf tower out and using the lights and seeing whether or not things work so what it means is that you really really have oops to fail gracefully um you need a backup plan in case things don't go well and I know this and the people running the golf tower know this which means they we together we sort of came for this really clever idea that we would send them data and if at ever we stopped sending them data you know it went more than five minutes without sending a update to them they would say oh something must be wrong and instead of the north wall being green and the west wall being red the west wall was green the north wall was red and they would just play animations on the building that looked like things were happening and everyone in Pittsburgh it would look just like the everything was happening normally and we would know oh no the wrong wall is the wrong color we need to go in there and we need to fix it and thankfully we didn't have to use this particular piece of software but it's really really nice to have so we have all of our pieces of software um they're all running out there and we ended up over the three-day period getting about 16 000 photos um Pittsburgh posts a lot of photos and of those we ended up with about 8 000 positive about 3 000 negative the average score was a 0.3 this was our happiest photo that we got over the three-day period apparently there are some people who write entire love letters to their wife on instagram um as public posts with all sorts of you know it was this huge huge letter about how wonderful she is and this was our saddest one which was actually the sort of heartbreaking story of a girl who suffers from depression but is coming out of it and everyone's sort of encouraging her in talking about their sadness which is remarkable because it worked really well and these are all of the instagram photos that we that we got over Pittsburgh the bigger the circle the more positive or negative it was the you know red ones are negative green ones are positive gray ones are neutral um to get a little closer here's where Pittsburgh people send instagram photos from you can see oh look Oakland is really popular downtown is really popular um you can actually sort of see the south side down there as a bar and you can actually get in really really close and say oh look there's stage ae there was a concert that was happening there were a bunch of photos from there you know there's a hockey game going on over here you know again there's the south side down there um there's that little strip on pen avenue with all the bars on it and you can sort of see where people take photos and it's really interesting to sort of look at all of these results across it and one of the interesting things that started happening is people started taking photos of the golf tower and talking about it it's really cool to see a project that you've worked on and all of a sudden have random people talking about it you know using the social media tools and sort of because I can't see the golf tower from where I live so I turn this thing on and the first thing that happened is I started going to twitter and looking for people talking about it so I could see photos of what was happening um the other really thing that happened is that people started trying to play with it um all of a sudden we started getting these like let's make the tower green today or I bet I can make the most positive comment and get to be the top score of this and all of a sudden all of these games started appearing around this thing sort of emergent it never occurred to me that this would happen but it's really really cool to sort of see what happens with the technology when you stick it out there another really good story is we were watching these results as they happened it was really interesting because all of a sudden Pittsburgh's mood dropped about 15 percent and it turns out every time the walking dead comes on everybody in Pittsburgh starts talking about death and dying and people getting shot and zombies and you can actually see this drop in the general mood of Pittsburgh at least on instagram and it was it was kind of cool because we started getting attention that you know the museum doesn't normally get we were an art museum we talk about sort of hoity hoity things sometimes we get into like the new york times um engadget doesn't tend to cover the Carnegie Museum of Art but this is but all of a sudden we're on engadget or we're on psfk um we also started getting local attention and it was really really it's you know every time the museum sort of puts on an exhibition they send out press releases the the newspapers and the tv stations start you know come and talk about it because we're a cultural institution they cover that but it was really cool that all of a sudden they started calling us back and saying hey what's the tower doing and you know when all of a sudden you're on the weather report you know that you're really really successful um and actually the way you know you're really really successful in Pittsburgh is when the mayor starts talking about you so if you're Slater Kenny you're really successful this means that we're sort of at that same scale um sort of why this matters as Brad was saying earlier we have this thing that we did and we're trying to we're the museum is moving in the 21st century and we're trying to figure out how to engage with people um how to engage with technologists and specifically and young people um you guys are sort of the audience that the museum is trying to reach and to do this we have to start we have to sort of learn as a museum how to think about digital technologies and social media and the internet um not just the stunts but as new ways to talk about art new ways to communicate with people um so you guys are the people that we're really interested in reaching and we're not always really good at and so what we're really looking for here particularly from an audience like you of people are interested in art people are interested in technology is is your ideas and your suggestions and your participation with us um museums have incredible data museums have incredible opportunities we're not always really good or really equipped to do cool things with it sometimes we can like this project sometimes we can't but we we we really want to work with people like you to take that information take that sort of cultural legacy because the museum's museum sort of holds all of this cultural information I mean we have paintings on the wall and we have sculptures but if you think about it those are just representations of information and and the museum's job is to hold those things in trust for the city of Pittsburgh which is you guys and that means that it's not just we hold the the painting we put it on a wall so that you can come and look at it it also means that we hold this data in trust so that people like you can get access to it and look at it and we're working to figure out sort of how to do that better with people like you and we're looking for suggestions and participation one last thing as a confession about this project I sort of it turns out that I lied about what the saddest photo in Pittsburgh was it turns out due to an algorithm uh bug the um the penguin's winning is the saddest thing that happened in Pittsburgh over that three-day period apparently when you talk about shots and you talk about things being over it scores really really badly um I just didn't want to be the person who told everyone that this is what makes Pittsburgh unhappy thank you all so much and in particular I wanted to thank Dave DeSello who took this gorgeous photo that we've been using for all of our talking about it he's a fantastic photographer and he's really generous in letting us use it I want to make sure to thank Brad for giving me this opportunity to sort of do a project like this which is such a it's so fun to do something like this in my hometown and sort of do something that this tower that I grew up watching and you know learning to read the weather from I get to play with it now um yeah and the rugby reality and Larry Walsh and Terry Gilliam who was the or 10 Gilliam who's the person who helped with the lights they were really really helpful also Justin and Code and Supply thank you so much for inviting me to give this talk and all of you guys who take pictures on the Instagram we couldn't do this without you does anyone have any questions great sure which I didn't before so I wanted to just mention some results from this so over the period of the golf the week of the golf tower project we experienced a 60% lift in web traffic to all of cmoa.org and all of our web properties we experienced a very large lift to the distant field exhibition page in particular and I wanted to talk about return on investment which for me is important as a marketer so we spent roughly five thousand dollars in development costs on this project a little bit more in advertising and our return on that was more than three hundred thousand dollars in earned media so overall fairly successful as you can imagine so and the fact that we got Dave Crowley from KDKA to quickly understand what the exhibition distant field was about was a was a huge win because his first question to me was now what's this Instagram thing you guys created I don't know how this interview is going to go so um so but but again thank you guys very much and I want to reiterate that if any of you want to come back and you are not members of the museum you want to come and visit um you know certainly we'll have some time now to go check it out um and I would encourage you especially to check out distant field after we've been talking about this it's right up in forum gallery up the stairs um but if you want to come back after today I I'm happy to provide any of you a free pass uh to to come back and see some more of the things that we have and if you have ideas the way David mentioned about ways to connect art to connect people with our art through technology or other means please let us know our next thing is is a project using oculus rift and I won't say any more about that but but we we started yesterday so we're really excited about that so we're always looking for these types of opportunities so thank you