Video summary

UTmessan 2020 - The Benefits of Open Data API

Main summary

Key takeaways

Technology

Technological Concepts & Product/Features Mentioned

Open Data: Definition & Licensing

  • Open data is framed as research/data made freely accessible, so knowledge can advance (with historical context tied to Robert K. Merton’s idea of freely accessible research results).
  • Modern open data is described as being licensed under Creative Commons Attribution 4.0 (CC BY 4.0), which allows people to:
    • copy, redistribute, remix, and transform
    • build upon the data (including commercially)
    • do so with attribution and a clear indication of changes

APIs as an Enabling Infrastructure

  • APIs are presented as a strong example of how open data becomes usable for developers and data scientists.
  • Example: an Icelandic “Center” / “swore Center” website (exact name unclear due to subtitles) that:
    • uses web scraping to harvest public data from multiple websites
    • exposes a well-structured API to make it easier for students/projects to fetch data
  • Reliability risk: if upstream data providers change their structure, the ingestion “chain” feeding the API can break.

Government-Led Digitalization

  • Iceland’s government initiative is mentioned as increasing public availability of government open data, also under Creative Commons licensing.

Use of External APIs in Projects

A transportation/climate cost project used:

  • Google APIs
    • for distance and trip time estimation (home ↔ work)
  • a license plate / car information API (subtitle indicates something like “API Stiers”)
    • to estimate car-related emissions
  • The talk emphasizes how quickly teams can move once APIs/data are accessible (e.g., “took only a few hours”).

Advanced Methods: Machine Learning Workflow

  • The talk argues that machine learning helps when problems are too complex to solve manually.
  • It also corrects a misconception: data science is not mostly algorithm work.
    • Much effort goes into:
      • accumulating data
      • data analysis
  • Key requirement highlighted: ML can’t progress without reasonable, sufficient data.

Reviews, Guides, Tutorials, and Actionable Takeaways (Hackathon “How-To”)

Recommendation: Join Idea-Based Hackathons

Suggested benefits include:

  • expanding your network
  • gaining access to mentors/experts
  • benefiting from Iceland’s startup/community culture
  • improving problem-solving via cross-background teams (market + societal perspective)
  • validating whether an idea needs iteration (the speaker notes five tries before a first win)

Specific Hackathons / Case Studies (Technology + Outcomes)

  1. Million Tonnes Challenge (Iceland Chamber of Commerce)

    • Goal: reduce 1 million tonnes of carbon emissions by 2030
    • Problem found: limited public data (only total carbon emissions from 2017)
    • Proposal: collect/publish real-time emission data so people can make informed decisions
    • Note: results were expected later that year (per the talk timeframe)

  2. Reykjavik City Hackathon

    • Goal: improve citizen living experience
    • Transportation challenge: students avoid buses due to timing reliability
    • Proposal: combine bus schedules with school timetables and other time-dependent institutions
    • Aim: build a virtual environment for bus systems to optimize capacity cost-effectively

  3. Climathon (Martis hosts / climate-focused hackathon)

    • Project: “Take the Test” (exact title unclear)
    • Purpose: compare commuting by car vs public transport, including:
      • estimated money saved per year
      • estimated travel time changes
      • carbon savings calculation using Google + car/emissions APIs
    • Reported results:
      • ~half a million ISK/year average car ownership cost
      • bus adds ~10 extra minutes per trip (or ~20 round trip)
      • ~1.4 metric tons of CO₂ saved annually (as stated)

  4. City Hackathon 2018 (non-winning, but progressed further)

    • Related to “clogging” / outdoor litter-picking
    • Goal: improve logistics for organizing litter “plotting”/mapping areas with a plugin/web interface showing which areas are clean
    • Reported impact at launch:
      • 1,000+ participants/marks across Iceland on a big day
      • ~10 tons of litter collected
      • coverage claimed over 17 square kilometers

Broader Analysis Points

Impact Statistics

  • More data-driven ideas are appearing:
    • >65% of ideas in Iceland hackathons are data-driven
  • But survival is limited:
    • >75% of data-driven projects do not survive
    • implying open data alone isn’t sufficient

Why Projects Fail (as stated)

  • Data “can’t survive on its own.”
  • Successful projects typically require additional domain support from other industries and collaboration to achieve real adoption.

Main Speakers / Sources (as Identifiable from Subtitles)

  • Speaker: presenter discussing open data in Iceland (name not provided in subtitles)
  • Referenced source: Robert K. Merton
    • historical framing of freely accessible research results
  • Institutions referenced:
    • Government of Iceland
    • Iceland Chamber of Commerce
    • Iceland environment agency
    • Reykjavik (city hackathon)
    • national land survey (mentioned as an open data source)

Original video