Hackathons: Bringing Different Industries Together Around Open Data

Hack·a·thon — an event, typically lasting several days, in which a large number of people meet to engage in collaborative computer programming.

The journey began shortly after starting my fellowship at the Vermont Center for Geographic Information (VCGI), the State of Vermont’s office of geospatial data and technology, when a friend reached out from HackVT. HackVT is a technology focused event and they were interested in using the State’s geospatial data in their 24 hour hackathon event. It proved to be a unique way to collaborate with local organization and technical talent to promote VCGI’s mission and the Open Geodata Portal. The event was also an opportunity to work towards one of VCGI’s objective for 2017: bringing together GIS professionals, data scientists, and software developers to learn from each other.

The goal of any hackathon is rapid prototyping to solve any unfulfilled need. Before my fellowship I had run several internal hackathons at my previous company so I was familiar with the structure and requirements. HackVT was being run by the Lake Champlain Chamber of Commerce, BTVIgnite and LaunchVT to promote tech for the City of Burlington.

The Challenge

This was the 7th year for HackVT with a 24-hour format, a panel of judges and monetary prizes for top 3 teams. From the start the HackVT committee recognized that Geographic Information Systems (GIS) and geospatial data was inherently more complex than previously used datasets. It was determined pre-event talks were needed to allow the community to ask questions, understand how to interact with APIs and file types. Several of us stepped forward to organize, develop sample apps and present about GIS.

Promotion

Hackathons don’t necessarily follow traditional marketing approaches for bringing in teams. Participants are usually tech savvy and often utilize specific digital services for their communities and conversations. In Burlington there are several Slack channels for developers based on different technical interests. We posted about the talks and HackVT in the various channels, announced HackVT details at Meetup events and used word of mouth to get the word out at various companies.

We also created a new Slack channel for the event itself. It was immensely helpful for logistics as well as providing a location to place links and announcements. Pages were added to the HackVT website as needed detailing talks, resources and datasets. Interestingly the HackVT Website traffic was predominantly from desktop (75%) even though the website was adaptive for mobile display. Once the dataset was announced the traffic jumped and plateaued. Half of the traffic was direct and there were spikes around pre-event talk dates.

Jenny Bower presenting about mapping

Pre-Event Plans

Because of the complexity of the geospatial data, multiple presentations were set up in the month leading to the hackathon. Overall four different Meetup Groups helped host six talks in three different locations. Attendance varied from 60+ to 14 people. The time line of events were the following:

  • Announcing of Hackathon Dates (Beginning of September ~ 1.5 months out) — Disseminated on the website and through word of mouth
  • Code Camp (Mid September~1 month) — Announcement of the data set and partnership with VCGI at the annual Code Camp along with the start of the Slack channel
  • Vermont Coders (several days after Code Camp) — First presentation with a Javascript Leaflet example using GIS
  • Code for BTV (2 weeks out) — Multiple presentations on the current state of GIS, example maps from QGIS and the future of GIS for a more general audience
  • Vermont Coders (1.5 weeks out) — Examples of how to use Javascript with ArcGIS, Leaflet and Mapbox to present web application maps
  • Burlington Data Science Meetup (same day as event above) — Intro to GIS, exploration of the State of Vermont’s Open Data Infrastructure and walk through with examples with Python Jupyter Notebooks
  • Code for BTV Hangout (3 days out) — Open session for anyone to come in and ask questions
Becky Grenier presenting on her app about the age of building in Burlington

Overall presentation attendance was great, and it quickly became obvious that many were generally interested in GIS with varying levels of experience. For me it was great to see that people understood overall the power and importance in exploring geospatial data. Curating the datasets and presenting clear documentation in advance on the HackVT website was vital to help teams identify which they wanted to use and what services were available well in advance of the hackathon.

The Event

Most registrations happened right before the deadline (which is consistent with my experience with hackathons) even after we extended it. In total there were 60 people making up 18 teams varying from high school students to seasoned professionals from Vermont and beyond. Doors opened at 3pm on Friday so participants could set up and get started.

Green Mountain Power provided a wonderful venue with open working areas, conference rooms and 24-hour access. Volunteers set up tables and chairs, set out name tags on work areas based on team sizes, put out banners and swag and set up registration.

All ready!

Next to the front door was registration, a help desk and the swag table. A presentation area with dual screens, podium and mic was for opening remarks, mini presentations during the hackathon, final team presentations and closing remarks. One conference room was set up for sleeping although people grabbed a nap wherever they could. The kitchen had a large cooler with drinks, snacks, tons of coffee and tea (you can never have enough coffee!) and served as the staging area for breakfast, lunch and dinners .

A table was designated for the two volunteer ESRI employees that came up from Boston to provide technical assistance. Power strips (the black lines from the celling in the pictures), LAN connections and logins to the powerful wifi network were provided as well.

Opening remarks were at 6pm to explain the schedule and rules. People trickled in for the first couple of hours to set up their monitors, laptops, a handful of gaming desktops and even a virtual reality setup. Teams were quick to get down to work.

Once everyone got underway it was quiet. The first couple of hours are usually planning and wireframes; tasks are divided, and work begins. All the tables were provided with whiteboards and markers, always helpful for mocking up ideas and setting priorities.

And many lists…

A presentation given by Chris Yager reviewed best practices on how to present their projects and helped the teams keep their talks short and to the point. It was instrumental in the quality of the final presentation as many teams were not experienced in public speaking.

In order to break up the night and wake everyone up there was a quickfire challenge at 1AM. These were not technical but rather challenges of physical dexterity. Competition was fierce as donated prizes were on the line, and it was a welcomed break from computer screens.

People took naps wherever they could find a good spot, or headed home if they lived close enough. At one point I looked into a conference room to see all the lights out and everyone in sleeping bags on the floor. Most of the volunteers had shifts so few were there for the whole night.

By mid day Saturday, teams were hitting crunch time. Presentation slot sign ups were opened 3 hours in advance. Each team would have 30 seconds to hook up their laptops and 3 minutes to present. Judges would then be able to ask questions about their projects. Not all teams ended up presenting; some were just there for the fun of the event.

At 3pm presentations began to a fully packed room. In total 13 teams demonstrated what they had accomplished.

Here is the list of teams with project names and short descriptions:

  • System 42 / ObserVR — Virtual Reality program exploring in 3D various points in Vermont with interactive map navigation
  • Emergent Vermont / Emergent VT — Using tile maps to show various landmarks, recreation areas and ski trails
  • Essex STEM / Revolutionary War History -Story Map from a local High School team depicting the history of the Revolutionary War in Vermont
  • Team TBD / What’s Next? — Dynamic filtering of potential events and activities based on criteria and with a voice activated Alexa interface and a mobile app
  • The Freedom Unit / LYM — Augmented Reality geolocation tagging viewed in place through a mobile app
  • HackHawks / Jot.VT — Mobile app returning nearest recreational pont and a mini-map navigation centered on your location
  • UVM CS Crew / Fish Vermont — Displays all the local fishing access points and launch types.
  • Magellan / Soiled — Contextualized soil, agricultural value and flood data for gardening applications
  • Fishar / FishTails — Gamification of fishing hotspots with location based check in to access points and fish recognition
  • $eal Team Rick$ / HistorVT — With dynamic date range criteria return local historical sites based on historical roadmarkers
  • Hackers From the NEK / Vermont Cow Talk — An automated cow-themed chatbot using voice recognition from Alexa to return interesting history, locations and google street view
  • Galenerds / Pioneer VT — 3D elevation interactive map depicting event, location and reviews from Google API and the ability to place geolocated shared flags
  • Mind Control / PopUpHistory — Proximity popup of historical location with check in points and background processing

After presentations ESRI announced their bonus prize to the Essex STEM team of free developer licenses for AcrGIS, closing remarks were made and everyone adjourned for food while voting was tallied.

HackVT Results

1st Place — $5,000

Team — HACKERS FROM THE NEK
App — VERMONT COW TALK
Let’s you ask a local cow(bot) for suggestions on what to do in their area.

2nd Place — $3,000

Team — MIND CONTROL
App — POPUP HISTORY
Notifies your smartphone when you are near an important landmark, historical site, or recreational area.

3rd Place — $1,000

Team — TEAM TBD
App — WHAT’S NEXT?
Uses voice recognition and AI to recommend and schedule your next activity based on what you've done, like to do, and where you are in VT.

Student Prize — $500

Team — UVM CS CREW
App— FISH VERMONT
Populates a map with fishing locations based on your search requests for type of fish, boat access, surface type and location.

The Experience

We truly had a blast during the event. It was great to watch people work and see the remarkable products made in such a short time. It also provided an opportunity to meet new people and make connections that otherwise would not have been possible.

Volunteers hard at work

The collaboration between VCGI and HackVT was a great learning experience. VCGI focused on pre-event presentation, documentation and datasets and HackVT did a great job with logistics, fundraising and promotion. It was a balanced mutually beneficial partnership that built new connections between multiple organizations while also enabling innovation.

Here are a few takeaways:

  • Slack — Slack was useful as a logistical tool coordinating talks, troubleshooting, pre-event planning and during the actual event, but was only lightly used by actual participants. The channel will remain active and will be utilized for the event next year, but the separate public help desk channel may not be needed. It was also useful for posting links, Github repositories, articles, announcements and thanking everyone at the end.
  • Help Desk — It is unusual to have a help desk for a hackathon. There were occasional questions throughout the event but it was never very busy. The help desk was for data problems, not coding questions.
I also put my art degree to use throughout the night by drawing on a white board
  • Pre-Hack Events — There was a lot of general interest in geospatial data and the events were often packed. We tried to gauge who was there for HackVT and it was often a small percentage, most were just interested in GIS. These events successfully provided materials supporting what seemed to be a large interest in geospatial data across a variety of communities.
  • Theme —The general theme of “Undiscovered Vermont” helped teams focus on datasets and building a compelling story about Vermont without constraining them. There were certain trends that emerged from the teams in which datasets they choose with recreational areas and historic road markers being at the top.
  • Final Products — It was not clear from the very beginning what kinds of projects would come out of the hackathon in both the subject and quality. We all knew that integrating geospatial data and technology could be complicated but teams were able to provide working code from virtual reality, fishing apps, exploration and story maps.
  • Collaboration — An unexpected benefit of the event was the connection of multiple communities in support of the hackathon. It brought together different state agencies, the Burlington civic hacking group, private companies, chambers of commerce, colleges and local government. What future collaboration may look like is not clear, but new channels were created with common goals and understanding. We also requested if anyone wanted to host their content on the VCGI Github we would welcome that opportunity.
  • Open Data Portal Exposure — A primary goal for VCGI was to expose a new audience to geospatial data available in the Open Geodata Portal and drive it’s usage. This goal was achieved, but it will be interesting to watch portal usage over the next couple of months to see if the effect was “sticky” or not.
  • Presentations — We were lucky to have first-class presenter Chris Yager talk to the teams about presenting. From talking to others the teams following his suggestions had much better presentations and elevated the overall project quality.
  • Professional Development — Every developer who went to a presentation got a little introduction, and those who actually participated in HackVT each learned the basics of how to use GIS data. In my conversations with some of the participants they talked about how it changed their perspective on their own work and undoubtedly will affect local companies and increasing value of local developers.

There is a lot to be gained through collaboration between public and private sector in the name of innovation. I hope to see some of what was made get put to good use and a continued expansed use of geospatial data in Vermont’s tech journey.

--

--

Kendall Fortney
Vermont Center for Geographic Information

Self-taught Data Scientist focused on Python, machine learning and Geospatial Data with degree in Art and years of experience in tech in Vermont.