SolarSPELL: Survey ETL + MySQL + Analysis

ETL
Python
SQL
Cleaned multi-country survey data, standardized schemas, and staged to MySQL for downstream analysis.

Problem

SolarSPELL needed a reproducible pipeline to prepare field survey results for analysis.

Data

Collected responses from multiple countries with varying schemas and encodings.

Approach

  • Normalized column names and coded categorical values
  • Loaded cleaned tables into MySQL using SQLAlchemy
  • Produced summary dashboards to explore regional trends

Results

Delivered a reusable ETL script and database that cut data prep time in half.

Visuals

Sentiment Analysis results of all free-text responses from a Survey

This was collected through ChatGPT’s API. All qualitative and free-text responses were collected and given a summary(theme) and sentiment score

Top Themes from Open-Ended Responses

Bar chart of the five most common themes explaining usefulness. “Resource availability” dominates, followed by affordability of technology, cost of books, cost of educational resources, and variety of tools.

CA Plots: Productivity Lift for Users

Slide showing boxplots for Conservation Agriculture (CA) plots. SolarSPELL users have higher and more consistent yields, with median productivity ~49.15 Kg/Are vs ~45.00 Kg/Are for non-users.

SP Plots: Consistent Gains Across Methods

Standard Practice (SP) plots show the same pattern: users’ median productivity ~38.00 Kg/Are vs ~30.46 Kg/Are for non-users. The note highlights similar gains observed for other crops grown with CA, reinforcing broad utility.

Knowledge Sharing: More Farmers Training Others

Simple bar chart comparing the share of farmers who trained peers. SolarSPELL users are more likely to teach others—~82% vs ~77% among non-users—underscoring community diffusion effects.

The “Wow” Stat: Amplified Community Impact

Users train far more people on average (≈103 individuals per farmer) than non-users (≈42). The slide frames this as ROI and scalability: each user becomes a local knowledge hub.

Implications & Next Steps

Summary slide: (1) Proven model—evidence of better agricultural outcomes; (2) Leadership—users become educators, enabling sustainable development; (3) Expansion—strong case to target women coffee farmers to reduce gender gaps in access to information.

Repo / Live