The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
It’s summer in the Southern Hemisphere, and families in this area are heading to the beach. One of the most amazing places to visit is Florianópolis, known for its vibrant culture and seaside charm.
To understand what makes the city’s Mercado Público so special, I built an end‑to‑end data pipeline: scraping Google Maps reviews with Python and Selenium, storing them in PostgreSQL, and categorizing them with Pandas. The insights are visualized in an interactive Tableau dashboard. It's also the first time I used another tool: Beekeeper!
This project blends data engineering, sentiment analysis, and business intelligence to highlight how people experience one of Brazil’s most beloved destinations.
You can explore my GitHub repo for more information.
Have a nice week! ♥️
Cool
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
I built an AI agent that handles customer operations — refunds, order lookups, and support tickets — with a human approval gate built into the workflow. The agent proposes the action, pauses, and waits for a human to approve before anything executes. The LLM never makes the final call. Role-based access and a full audit trail are enforced in code, not prompts. Deployed live on Azure Container Apps.
The pause for human sign off keeps refunds safe while still offloading the lookup work to the agent.
Standard municipal data is misleading.
If you want to find 15-Minute Cities within Greece's urban landscape, you have to rely on more than just city-wide averages.
I've been creating a containerized PostGIS spatial data warehouse to identify undervalued, structurally neglected and high-potential urban micro-fragments.
The data includes:
6,138 official ELSTAT communes
12,505 objective government pricing polygons
132,000+ OpenStreetMap amenities
Fusing different spatial and flat tabular data together presented challenges. Last night, I published an article on Medium documenting the process of building this spatial ETL pipeline. I cover the architectural wins, and of course the bugs encountered along the way.
These include:
Spatial Sliver Confusion: How a 150m² geographic sliver falsely inherited the population of a 20km² city, and why I had to build a Two-Tier Data Model to fix it.
CRS Misuse: How a projection mismatch between EPSG:3857 and EPSG:4326 caused the mapping software to panic and collapse into a blank screen.
Missing Spatial Index: Why dropping a functional GIST ((geometry::geography)) index turned a 2-second query into an hour-long frozen process.
Analytical Reward: Running an initial and experimental spatial linear regression in SQL, showcasing a €8.51/m² property value increase for every additional amenity per km².
If you are into Data Engineering, GIS or Spatial Engineering in general, check out the full technical breakdown here:
🔗 https://medium.com/@kchantsaris_84495/designing-a-spatial-etl-pipeline-in-postgis-72565b53aeb8
An AI-powered operations assistant designed to help property management teams quickly find answers from their property, maintenance, and work-order data.
Instead of manually searching through spreadsheets and work-order records, a property manager can ask questions in natural language and get structured answers based on the underlying operational data.
The system can identify unresolved maintenance requests, high-priority issues, properties with the most open work orders, and work orders that have been open the longest. Relevant supporting records are displayed alongside each answer.
Workflow:
Property & work-order data → Natural-language question → Data retrieval → AI interpretation → Answer + supporting records
Key features:
AI property operations assistant
Natural-language business data queries
Maintenance and work-order analysis
Open issue identification
High-priority issue detection
Property-level operations analysis
Supporting records for AI answers
Structured operational database
Local/offline fallback
Built with: Python, Flask, SQLite, HTML, CSS, JavaScript, SQL-based data retrieval, and AI-assisted question interpretation.
This project demonstrates how AI can be applied to automate information retrieval and support day-to-day property management operations.
