 Hey everyone. Thanks for coming. I'm Chris Ritzo and I'm with the Measurement Lab Project and I'm our program management and community lead. A new title for me since we just shifted and became a fiscal sponsor program at Co4Science. And so we're really happy about that and being a part of that family now. My goal with this talk is to expose you to Measurement Lab and our data set and see what people are doing with it and see what you might be able to do with it if you're interested, but also to give kind of an overview of the project. So we are an open data platform that provides a measurement service, an internet service measurement platform. So we host measurement tests on servers around the world and then people can run tests against them to understand the characteristics of their internet connections. So their speed, latency, you know how fast is your internet essentially. And then all of the data goes into a public domain data set that's available openly to anyone and a big part of my job is supporting people who want to use that data. The GitHub repo, well our organization is up here on GitHub so you should check those out. And there's a lot of repos in there so if you're getting into the code just come up to me and we'll talk more about what you're looking for. So a real quick quick start if you're in the room and you want to just start diving into ideas, code, what's there. I wanted to put this slide up here to highlight a couple quick things. We have a visualization website that aggregates our data by location in different locations around the world and a couple little screenshots here. Illustrate that. Comparing ISPs in this case in Portland over the last six or eight months on this little graph and then the right hand or this side under the data it shows that you can access that data via the NAPI. So that's a quick way to get started looking at our data if you want to start if you're building your own applications. Just start using that REST API with or download raw data. And then of course if you want to dig in more deeply and look at individual test rows for a specific location or time period we have all of our data stored in BigQuery and there's a quick start guide that we've written that helps you get started with that. And of course the support address if you want to contact me or some of my colleagues to get help. So our mission is here on this slide measure the internet save the data and make it universally accessible and useful. This is quite huge and aspirational but it keeps us moving toward a real that really aspirational goal. So I'm going to talk about some of how we measure it how we save the data and then we'll switch and look at the accessible and universally accessible and useful part because that's I think the bigger thing that we can't do by ourselves. And this slide kicks off this part of the talk because our community is made up of a lot of different types of communities from companies who donate money or server space and transit to our servers to the folks who design researcher research experiments that are run on the platform to the policy space and academics who want to use this data as a secondary resource in their in their work. So tons of different types of audiences and this is a big challenge for us to try to you know balance meeting all of their needs. We run a bunch of different types of experiments or tests and these are some of the ones that we've hosted over the years. The most highly used is this network diagnostic tool from that was originally built by internet to and it's a speed test. It's you know measures much like some of the commercial speed tests you might have encountered like speed test.net or Oocla. But all that data is openly available and all the code is open so people can integrate that code into a web site or software or hardware product. These are all tests that have come from the academic community primarily. So folks studying the internet about ten years ago started to think well we need this global platform to try to understand how to make the internet better and that's how MLab eventually came to be. So we have quite a number of tests that originate through the computer science and network measurement internet measurement community at different universities. Over the last ten years we've grown from and this is sort of a platform growth slide to show sort of how long we've been in existence and where some of the milestones of where we've placed servers over time. We have servers in lots of places. About 130 locations, 500 plus servers not in all areas of the world. So we're working on that. But you can see we've got a pretty heavy North American and Europe but a smattering across the rest of the world. All this means that users will be directed to the test based on their geography. So if you're in South Africa you'll be directed to our servers in South Africa and unless you use a tool that lets you select that. Color in this case is a bit dated. Excuse me, sorry. Yeah, color. So these dots essentially come from the map on our website. Red are some servers that are currently offline or when I took this snapshot. And then because each of the servers hosts a series of experiments some of them are fully available and some were not at this time. So that's the different shades of green there. We do all this work through a variety of partnerships and so different companies can and organizations can contribute in a variety of ways and this is just a smattering of the different logos from our site. We get a lot of corporate funding and we're working on sort of diversifying that. But much of our hosting already represents that diversity from the global research and education network of which some of these folks represent. Now the volume of data is quite large. So we currently get about 2 million speed test measurements a day from the network diagnostic tool and we hit a milestone two years ago with a billion rows of data. So it's quite large and that doesn't include any of the other tests that we run. So there's a rich, rich data set for researchers and advocates to use to understand internet service in a variety of ways. I am going to get to the data. Just wanted to call that out and what some of the people are using it for. The real spike in volume over the last three years came when the Google search team integrated NDT into a little widget. When you search how fast is my internet or internet speed test, Google provides this little widget where you can run our speed test. That data goes to MLAM and provides a service to anybody who's doing that search. So that's that huge spike and we have other integrations that have also helped with getting test volume. So I'm going to show some of those but Fingbox, if you've used the Fing app, that is a measurement of network quality, speed. They have a little device that you can place in your home that runs a speed test every so often and gives that information to their users. MLAM also has a Chrome extension for Google Chrome and if you'd like to run our test and save the results locally, that's a quick and easy way for you to gather data about your own connections. So those are more software and website type things but some web integrations are actually quite interesting. This is a screenshot from the Canadian Internet Registry Authority and they have built this internet performance test portal that integrates our tests and some other types of network measurements like DNSSEC and things like this and then they've mapped the data in Canada across, you know, in little hexagons down there on the map. This is a really great partnership because not only did they integrate our tests but they also host and sponsor servers in Canada. So this is a really good example of the type of partnerships that help us do what we do. So that starts to get into the last part which is how do we make it universally accessible and useful and we can't do that alone. It's a huge, huge goal. So what we do that is also through partnerships and supporting developer communities who want to use our infrastructure as a service and provide the data to their projects or analyze it in some way. So how is the data used is a good segue to this because it gets to the end point. What are people doing with this data? And this is a link to our publications page on our website where we try to list that some of the different types of things that people are doing. And so from left to right we have an academic paper that was recently published. It was just using our data to estimate household speeds and tiers of service from the University of New South Wales and some academics who are working there and they published a nice methodology. We did a paper with Alliance for Affordable Internet last year where we took our NDT data and segmented it by mobile and non-mobile tests and then kind of looked at the aggregate for a short exploratory report with them on mobile broadband service in various countries. And then the third example is an example of advocacy organization that's looking at media ownership and consolidation issues in Venezuela. And for the last year or year and a half we've given them some data exports to support their journal, their advocacy. This particular one was around their elections, their special elections last year. So they were looking at public media and how much did independent media, independent journalists, how much of a voice were they getting and they were doing some traditional type of election monitoring as well sending people to polling places and seeing if they were asked to stop taking pictures or something. And then they also looked at internet speeds in aggregate for the different areas in Venezuela around those election dates. We don't often go out and seek to find these these publications, but some most of the time people just tell us about them. We're not doing a ton of data mining. This isn't on that where is an area where we could we could expand. Another project that actually brought me out here last week was this was a project that we're doing through the Institute for Museum and Library Services where we're taking our client test code and putting them on small computers, placing them in a library and those computers are automated to run our tests at regular randomized intervals. And that data will then go to a digital visualization platform that we're building for that project. It's a really exciting connection to the on the ground use of why why will we have this why will we do this to help public institutions understand the types of speeds that they're getting. So now we'll get into a few more web integrations. And this is kind of illustrating a few mapping initiatives. So combination of like a website that runs our speed test and then aggregates the results on a map. In this case, and in some cases asks communities to answer a survey set of survey questions. So these are targeted more at cities, regional governments that are, you know, wanting to understand the state of broadband in their communities with an open data set and open source code. And so what they've done and we've done in partnership with many cities and groups or to build, you know, build this website or help them work on it to build kind of a community engagement tool and like planning an analysis tool. So we partnered with this year and last year with the Center for Royal Pennsylvania at Penn State University, the Institute for the Local Self-Reliance. And they are building a report and have published that back, shared that back to the state legislature in Pennsylvania. And we've done some other smaller scale like pilot studies in partnership with Seattle and some other rural communities in Eastern Washington state. And then this is another example that was built not by us, but was supported by MLamb. Louisville Kentucky did a website like this that aggregated our data by zip code. The same idea, take a test, answer a survey, give the community data aggregated by zip code in order to really understand like where they have good service in their community and where they don't. And they leveraged that to decide where to build out fiber to areas that were underserved. And so this was a really big digital inclusion win that, and this code is now being taken up by the Tech Association of Oregon. And they did a hackathon last two weeks ago, where they're trying to build a national model out of this code to aggregate and share, make it something that many communities can use. So that was a great opportunity there. But the same sort of thing, you know, they're running a test, doing a survey, looking at the results and making decisions about it. One other example in the same vein that we published and worked with the Merit Research and Education Network and Michigan State University this year. The same kind of idea, but they hooked in a homework gap component where they built the science and then they are engaging with a sampling of K 12 districts across the state to take a home and assignment and run a speed test from their home to understand like, where the kids in the in their homes, you know, what is their service like? And a couple other screenshots there. And then this is another example that's looking at the large scale of the US. So this is a project that's soon to be published from our friends at Open Technology Institute, New America Foundation in DC, where they're looking at speeds from measurement lab in comparison and difference between our data and the FCC's data that ISPs submit about where they provide service and how fast it is it and what type of speeds is it. And then the last example is sort of a it's a mix. So right now, this is a great tool for aggregating data of various types, including M labs data called the internet is infrastructure.org and the I three connectivity explorer. And so the developer of this product had this this site. It has done lots of grad GIS work to bring in many different types of shape areas and to bring in data from the Census API and the ProPublica API, lots of different data sources to understand and look at public data alongside of M lab data and other data sources to really get to interesting planning outcomes for for communities that that that's his goal. So it looks at count, you can look at counties, you can look at sub county regions, you can look at tribal communities, you can look at micro urban communities and shapes and areas is a really powerful tool. And so what is all this mean for us is the next bit. M lab supported quite a number of these initiatives. And that's why we're here talking about our data and with you because, you know, we, we have supported lots of different individual initiatives. But to really scale and grow the power of our data in communities. Thank you. We are wanting to bring all these things together into a unified set of community tools that give communities and organizations options for doing what they need to do for their for their communities. And so we're planning a developer and stakeholders convening to scope and build that project. And if you'd be interested in that, we can talk more about that. But that's kind of the reason we I wanted to share kind of the ecosystem of different tools. Just dipping into the possibilities of what we can do with our data and to give a few examples. And now just stop talking and see what you have to say. I'd love to take some questions.