Data preparation & analysis
Making sensor data usable
Data Analyst
EdgeAi Labs
Python and Pandas workflows for sensor-data preparation, with findings explained across technical and non-technical audiences.
- My part
- Built Python and Pandas workflows to clean, validate and normalize sensor data, and explained findings to different audiences.
- What I delivered
- Prepared data for model development, analyzed data quality and documented findings.
At EdgeAi Labs, I built workflows to clean, validate and normalize sensor data for model development, using Python and Pandas. I analyzed data quality and usage patterns and documented findings for technical and non-technical stakeholders.
Before the model gets the data
There is useful work before a model receives an input. What does a field mean? Are the values consistent? What needs to be checked before a pattern can be interpreted? Those questions help explain why data preparation belongs in the account, rather than disappearing behind the name of a model or framework.
Make the finding understandable
A technical record should let another person understand what was done to the data. An explanation for a different audience needs to make the finding usable without assuming they worked through the same analysis.
Preparing data and helping people understand what it shows are connected parts of the work. I’ve included a small synthetic example elsewhere in this space to make the questions tangible without using employer data.