Fighting Air Inequality: Bridge the Gap between Programming and Thinking
Colleen Rosales, Strategic Partnerships Director, OpenAQ Inc.
This page requires that JavaScript be enabled in your browser.
Learn how »
- Organize, visualize and model large amounts of data in less time
- Easily develop GUI tools and share them with colleagues
- Approach data analysis in an intuitive way
- Air quality and environmental scientists
- Atmospheric and indoor air chemists
- Data analysts
Challenge
Air pollution is the second leading cause of premature mortality, and clean air is not distributed equally around the planet or within countries. OpenAQ aggregates publicly available real-time air quality data to monitor and solve air inequality.
Colleen Rosales, strategic partnerships director at OpenAQ, gathers data from a variety of instruments to analyze. Some take measurements in real time, but the frequency can vary from every minute to every second. Other instruments use filter-based methods in which air passes through a filter for a certain period of time. The filter is collected and analyzed for particles later.
Rosales wanted to visualize and model the data from all sources to see trends and compare which types of measurements performed better for different elements. Also, because the frequency of data being collected was increasing, the analysis was taking longer. She needed a tool to align the data collected from different instruments, and she needed that tool to work faster.
Solution
Rosales rewrote an existing data analysis program in Wolfram Language and created a graphical user interface for processing. Intuitive code and natural language "allowed me to combine my way of thinking into how a computer would think," she said.
She used the built-in function Dataset because it makes the data both human readable and computable. With human-readable data, Rosales can easily do sanity checks for different pollutant elements or different geographic locations. For example, she can easily check that in urban locations, the data shows certain elements from brake and tire wear.
In addition to built-in functions, Rosales used the extensible nature of Wolfram Language. Functions in the Wolfram Function Repository let her see more correlations, and she can also customize the functions further to compare different kinds of measurements clearly.
Benefits
Rosales said that Wolfram Language "allows you to think in many ways on how to approach your data, or how to analyze your data." This saves development time and increases flexibility because you can think about your program "in a way that is closer to how you'd normally think about it in your head."
