owned this note   changed 2 years ago  
Published Linked with GitHub

Sense-Making Journal group 1

MDEF: Measuring the world / A world in data activity report.

Team:
Amanda
Claudia
Ahmed
Caglar
Korbinian

Journal Index

From objectives to the hypothesis

Brainstorming

Image Not Showing Possible Reasons
  • The image file may be corrupted
  • The server hosting the image is unavailable
  • The image path is incorrect
  • The image format is not supported
Learn More →

Project Goals

Objective

We want a better work-life balance

A lot of questions

  • Do I spend enough time in nature?
  • Is my work fulfilling?
  • Is my identity / life shaped mostly by my work?
  • Are our salaries high enough?
  • Do we get health care?
  • Do we have recovery time at the end of a work day?
  • Do we have time to connect with people in a meaningful way?
  • Do we have time and energy to do what we love?
  • Do we feel we are free to choose what we do?
  • Do we have enough time for our mental health?

Pick one question and make a statement out if it:
Do we have time and energy to do what we love?

Hypothesis

People lack the time and energy to pursue their passions outside of work.

From Hypothesis to Data

Tools selection

Arduino

First Experiments

We started gather data to understand the difference between natural and artificial light. We used a a resistor of 1K.
March the 9th, rainy and cloudy day in Poblenou, Barcelona - We still could spot significant differences from outdoor and indoor:

  • Artificial light in MDEF class: 2000 K
  • Artificial light in the hall of IaaC: 1000 K
  • Natural light outdoor(cloudy): 4095 K (max)

Data Capturing Strategy

We used two different strategies to capture data:

  • with a sensor
  • from a data set (Twitter and Wi-fi).

Materials Needed

Adafruit Feather ESP32
LDR Sensor
Laptop
ArduSpreadsheet
IFTTT
Orange3
Twitter account

Detail Setup Instructions

  1. With Arduspreadsheets: sensor data from the serial collection to a .CSV file
  2. Gather the data via Wi-fi connection
  3. Generate more info through database - Orange3: Text (add-on) - Twitter and sentiment analysis

ORANGE3

First Trial: World Happiness Report

First trial we tried to upload data from the World Happiness Report 2022, the range was too wide, too many information and detaled classification. It is based on too many socio-economic aspects, that's why we decided not to base our research on it.

Second Trial: OECD.Stat

Then we decided to narrow down the research field and get the data drom OECD.Stat, a survey that gathers data about the percentage of adults who reported that over the last 12 months it has been difficult for them to fulfill their family responsibilities because of the amount of time they spent at work.

Third trial: Twitter

Collecting data from Twitter throughout Orange3 - SENTIMENT ANALYSIS
Query word list: over time; over working
Language: english
Tweets: 1000

We created a word cloud out of it

And then analyized the emotions throughout a Box Plot

LDR Sensor Data

To collect and log the LDR sensor data we employed a couple of strategies. First, we downloaded Arduspreadsheets and were able to log the LDR data in a .csv file. We took readings of the resistor value every 10 seconds and each reading was stored in a spreadsheet with a corresponding timestamp.

After this we realized we wanted the device to be mobile so it could be carried with a person to monitor the light levels they encounter throughout the day. For this we first tried to connect via bluetooth to a cell phone to collect the data. We made the bluetooth connection and could see the data but couldn't find a good way to save the data. After this we went with the option of sending the data over WiFi to an online spreadsheet. To make this happen we used If This Then That (IFTTT) to connect to Webhooks to publish the data over Wifi. After this we were able to connect the ESP32 to a phone hotspot and the device was mobile! We tested out different light levels throughout the IAAC building and outsice in the sun.

Tips

Tutorial for publishing sensor readings to Google using IFTTT

Arduino Serial to Spreadsheet

Pick the right sensor/tool for the application. In the end we didn't find the light sensor readings as useful for supporting our research for the time we had.

It will take WAY more time than you imagine to get everything working. Make extra allowances for time.

Data Capture

Data Summary

From OECD Stats

Below upper secondary education Upper secondary and post-secondary non-tertiary education Tertiary education All levels of education
Italy 56 58 48 55
Spain 45 58 56 52
Türkiye 88 82 72 81
United States 58 47 55 52


https://stats.oecd.org/

Data Summary
Project Title LDR Webhooks Events
Capture Start 03-10-2023
Capture End 03-10-2023
Original Data Format Light Sensor Readings
Submitted format CSV file
Total Data Points 400+
Number of datasets 1
Data Repository https://docs.google.com/spreadsheets/d/1YFaE637PvN5C3pGN2pU6fEf1k1q3pGcjHVHkii8CwdA/edit?usp=sharing
Data Summary
Project Title LDR Readings Day 1
Capture Start 03-9-2023
Capture End 03-9-2023
Original Data Format Light Sensor Readings
Submitted format CSV file
Total Data Points 71
Number of datasets 1
Data Repository https://github.com/agjarv/mdef-a-world-in-data/blob/main/ldr_sensor_manual2.csv

Data Insights

A hypothesis may be testable, but even that isn’t enough for it to be a scientific hypothesis. In addition, it must be possible to show that the hypothesis is false if it really is false. Proving it's true it will require testing all possible combinations, that's hard, maybe impossible.

Reflections

  • Our Arduino tests could have been more successful if we had more time to dedicate to locally gather data around students.
  • It could have been better if we selected our hypotesis more specific, we started from a way too broad and wide range. We needed to be more focused since the beginning.
  • When analazying data with Orange, is important to understand the range or data you're feeding to it. If it's too wide, you may loose the focus easily and not get the results you're hoping for.
  • Lower our expectations on finding amazing data.
  • In the reseach, we should have designed a process and be more organized about the data gathering. We lost our main goal while we were checking the data.
  • Trying many different techniques in one day was at the same time interesting but confusing and brought a little bit of chaos in the process.
  • We found interesting data on world happiness based on socioeconomic indices that we didn't get to use as much. For going further it could be interesting to use a sensor to monitor how much sunlight a group of people get per day and compare it with the happiness index.
Select a repo