Welcome to Hacking Measurement
We’ll get oriented and be introduced to the technology for organizing the class and the BIDS collaborative.
Reading / Tech
Please have a quick look at the Jekyll Docs. We’ll also use either Slack or Gitter to communicate about things – I haven’t decided yet. If you’d like to have a say, read up on them ahead of time!
Tentative plan
- Introducing the class – who’s in the room?
- The short history of the BIDS Collaborative
- Development Engineering: the potential impact of sensors, mobile, and satellite data
- About the class website, and expectations for contributing
- Introducing the project orientation
- How to contribute to the website with Jekyll (it’s part of the grade!)
- Looking ahead to basic sensor hacking (Javier Rosa)
Notes from Class
Dav Clark
email: davclark@b……y.edu
BIDS
- Website: http://bids.berkeley.edu/
- Location: 190 Doe Library - accessed to your immediate left when entering the library’s main entrance.
Root Class URLs
People’s Areas of Interest
- Education
- Entrepreneur
- Economic Development
- Environment
- Wellness
Temina Madon
- Works with CEGA.
- Data analytics in development. Looking for projects and students to pair.
- Starting ‘pseudo-discipline’ of development engineering.
- Starting a new journal (Development Engineering with Elsevier): design, measurement, etc.
- One goal: learn from development programs’ failures (e.g. PlayPump).
- Up to now: utilize popular development methodology of randomized control trials (RCTs)—experiments in development economics.
- New goal: “revolutionize measurement.” Current methods are costly, can be inaccurate (recording error, self-reporting errors, general bias), are infrequent, etc.
- Projects of interest:
- The Goldilocks Project
- Estimating agricultural yields
- Identify farmers’ plots using satellite data (high-resolution needed for half-hectare plots)
- David Lobell and Marshall Burke
- Use data from Skybox
- Sensors
- Do people use cookstoves?
- Response and cognitive biases appear to be pretty high here:
- 99% of respondents claim to use these cookstoves 3 times/day, every day.
- Sensing data suggest considerably lower levels of use (approximately half of reported levels).
- Other applications: filters, latrines, and power.
- Governments (might) want to know: does investment in technology yield results?
- Smartphones
- Premise Data
- Push tasks to agents (e.g. pay agents for images of households within 1km).
- Reliability of the process? Not technically crowdsourcing.
- Passive Measurement
- GridWatch
- Crowdsourcing
- The Goldilocks Project
BIDS Collaborative
- BIDS Collaborative’s Major goal: amplify on-campus collaboration in data science at Berkeley.
- Past projects:
- Visualization projects using energy data (D3)
- Supply chain management
- Market analytics for underwear startup
- DeStress: cognitive science
- High-frequency trading
- Currently building a set of courses: (Hacking Measurement, Diversity, …)
General Course Stuff
- Course website:
- Taking notes using etherpad
- We’ll follow the course website pretty closely. See the schedule.
- The syllabus is a work in progress.
- The course website uses Pages with GitHub. Here is the repo.
- To update the notes on GitHub: go to the lecture in “…/syllabus/_posts/” an update (using MarkDown).
- Reason for the waitlist: force everyone to build/join teams early. Teams then get registration codes. And teams need leads.
- People are welcome to introduce their own projects.
- Dave will distribute a list of projects soon.
- The next few weeks:
- Take a shallow-to-medium dive into the tools.
- Start with sensors (Arduinos, …).
- Satellite imagery.
- Utility data.
- Use familiar software/languages. Projects will generally determine the tools.
- Next Tuesday: crash course for GitHub. (190 Doe Library)
To Do:
- Can we get slides from Temina? Person 2 for today’s note-taking can follow up.
- Dav needs GitHub ids from each person. Leave your id (and possibly name) on the course’s repository.
- Optional: Buy some Arduino stuff. Dav can help put together a nice shopping cart.