Environmental and Agriculture Sensors and Data


#1

Environmental Sensors and Data Collection

Date: 23/04/2017
Attendees: André, Stacy K, Ali, Coco, Shannan H. Marina de Freitos , Leonardo Sehn, Shan He, Nur Akbar Arufatullah, Mario Behling, Gustavo Pereyrairujo, Andrew Thalor, Lawrence Ndero(sp?), Juan Keymer, Greg A., Dorn Cox

Overview of topic (3-6 sentences): We discussed a variety of topics about data collection tools, sensor types, data formats and how to know what to measure. Our main action item was to create a collection of useful tools, not just a list of links, but actual short videos describing each tool in its use and design.

Notes:

At the start Coco, Shan He, and one other (who was it?) split off to talk about more sustainable design and natural construction topics guided by the concept that the original open hardware was traditional knowledge that was passed down from generation to generation. (are there notes about this discussion? It was never brought back to the table at the end of the day)

The rest of the group was interested in discussing electronics hardware, sensors, and data.
Our discussion revolved around several central issues.

  • How to Measure.
  • What to Measure.
  • How to Store the data.

How to Measure.

We talked about validating sensor data and how important that is. Dorn Cox is working on a tool for Co-Deployment of sensors. Co-Deployment means installing an existing standardized sensor next to a new one to make sure the readings are the same. But this tool can be applied to non-technical data collection too. Non-Technical manual practices too. That’s one of the major challenges of building new sensors, is validating them and showing that they data they collect is accurate.

Just giving a citizen scientist with a project a sensor and a datalogger is not a way to get good data. Deployment requires some education and expertise as well.

It was discussed that it would be super useful to have a collection of resources. Not just a list of links, but short 10 minutes videos of each sensor or data logger explaining how it works.

We also talked briefly about note taking and getting metadata about the sensor deployment site. How, if it is a citizen science project, creating a detailed questionnaire gets better results than just saying “take a notes about sensor deployment”. Scientists have different note taking habits than non-scientists.

What to Measure

We talked about what different kinds of data is good for. Not all data is good for science, but sometimes you don’t need super precise data to fuel activism or protest. How much data is needed? Is it better to collect tons of data first and look for correlations, or ask a specific question and collect data to answer that question? What inherent biases are in each of these tactics? What platforms are the most useful for measuring and collecting data? If these systems were modular in structure, it would be easier to take pieces to re-use in different projects. EnviroDIY, ModelMyWatershed and Kibana (spelling?) are resources that were mentioned.

We talked about finding common indicators in the environment to help choose what to measure for monitoring projects. And what are alternatives to a sensor based approach.

Also mentioned as a helpful tool: Looking what types of materials there are regulations for (ie. lead, arsenic, nitrogen) and measuring those.

How to Store Data

We talked about data storage formats, and Shannon talked about the Observational Data Model format (ODM), which is an open standard that includes metadata about the sensor and deployment with the dataset. It has benefits such as making data comparable across sensors and data formats, etc. The Metadata (location, sensor type, etc) of a dataset is just as import as the data itself. Shannon is working on simplifying the process of building the metadata of the ODM format for the Mayfly logger for citizen scientists. ODM also is valid for non-technical data types as well.

ODM is part of the Globe.gov protocol for citizen science data to be accepted. There are standardizations across data collections and sensors. The Godan (??this is wrong) Initiative is working on standardizing data across devices as well.

Actionable Items

  • Online tools to upload data to a server
  • How to find out what to measure for monitoring
  • The creation of standards across data formats (or letting GOSH’s voice be heard in this process that is already happening)
  • Making a collection of available hardware and software projects (with a short video explaining how the work/what they’re for)
  • Creating API’s for environmental academic model (??)

If I forgot anything, or you can fill in some of the question marks, please edit this document! I did the best I could.


#2

Thank you for this excellent wrap up. Sharing random notes, that I have made in this session as well.

Dorn: JSON

  • Data Structure, Formats
  • Project, code deployment - drivers for sensors, validate, pair up
  • Democratize sensors in a field
  • Observational based
  • Universal observation apps (different protocols)

Shannon

  • Observational data model in the US: Makes it comparable
  • Collecting info about sensor (meta data) as well collected data
  • Put it in a database

Mario

  • database JSON/Elastic Search, loklak

Juan

  • What do you measure?
  • Juan-Pedro: Case dependent
  • Shannon: Question of quality

Mario

  • We need any data and as much as we can get to ask new questions tomorrow
  • deploy any sensor possible
  • teach AI level of validation (e.g. show me only

Andrew

  • The question is not what do you want not
  • Political question not technical question

Nur Akbar

  • Communication between sensors and controllers to LAN
  • Real-time controlling

Juan

  • How to use DIY sensors?
  • Data-sharing

Dorn:

Greg

Ryan:

  • How to make tools general and modular

Gustavo

  • If you know what you want to measure, but you want to use Open Hardware

Greg:

  • No “One” platform for everything. But: How can we make libraries more effectively?

Dorn

  • agricultural environmental models
  • nitrogen analysis for example

Mario

  • not one solution for all, but services for specific services
  • Kibana

Ryan

  • We need a list of tools

Greg

  • 10 minute video of tools/hardware

Kina

  • What is the data are gonna be used for

Andrew

  • Text files