From Petroleum Geology to GIS: Building My Skills in Public
For 13 years, I worked as a sedimentologist and stratigrapher for PDVSA in Venezuela and Ecuador. I mapped rock formations, studied reservoirs, and worked with spatial data every day — just not in GIS software. Since 2022, I have been retraining, step by step, to move into GIS work. This post is about how I am doing that, and why I decided to show the process in public instead of just listing skills on a CV.
Why show the evidence, not just the list
Anyone can write “GIS skills” on a resume. That line does not tell a recruiter much. It does not show how someone works, what they actually did, or how careful they were.
So I built a small website: a GIS Professional Competency Report. It is organized around 13 core GIS skills, like data cleaning, projections, topology, and cartography. For each one, I only publish a page once I have real, checked work behind it. Right now, two of the 13 are done. The other eleven are still open, and the site says so clearly. I would rather show a short, honest list than a long one that overstates what I can do.
Each finished skill has two parts. A short page explains what I can do and what I learned. A longer “technical evidence” page shows the actual commands, SQL queries, screenshots, and results behind it — including the mistakes.
Skill 1: Reading a new dataset before touching it
The first skill was about inspecting a dataset before doing anything with it. I took a CSV file with 200 records of sites in the Netherlands and checked it the way you should check any new dataset: What encoding does it use? Does it have real coordinates? Is there a defined coordinate system, or do I need to assign one?
I found that 8 of the 200 records had no coordinates at all, so they could never become points on a map. The other 192 had valid longitude and latitude values, but no coordinate system was defined in the source file. So I could not “prove” which system the numbers used — I could only assign a reasonable one (WGS 84 / EPSG:4326) based on what the values looked like, and I said that clearly instead of pretending it was certain.
Then I loaded the data into a PostGIS database, without changing or cleaning anything yet, and double-checked every number: record counts, how many points had valid geometry, how many did not. Everything matched. That match is the actual proof, not just my word for it.
Skill 2: Cleaning the mess, and admitting the mistakes
The second skill was cleaning that same dataset. This is the part I am most proud of, because it did not go perfectly — and I decided to show that instead of hiding it.
Some examples of what I fixed:
- One field had 26 different-looking values for what should have been 5 categories, because of typos, translations, and extra spaces.
- Dates were written in four different formats mixed in the same column.
- 21 rows had duplicate ID numbers, which I fixed using a reliable internal database key instead of guessing.
- Mixed units (square feet and square meters) needed converting into one consistent unit.
I also made real mistakes, and I kept them in the report on purpose. At one point, a bad conversion rule accidentally wiped out an entire price column, turning every value into empty (NULL). I could fix it only because I had kept an untouched copy of the original data — which is exactly why you always keep one. In another case, I completely forgot to clean a field called “area” and only noticed after I thought the whole task was finished. I went back and fixed it, and I wrote that down too.
I think this matters more than it sounds. In real GIS or data work, you will make mistakes. What matters is whether you can catch them, understand why they happened, and fix them without losing data. That is a skill on its own, and it is hard to prove just by saying “I am careful.”
What is next
Eleven skills are still open: projections and datums, digitizing, topology, spatial analysis, raster basics, cartography, database management, and a few more. I will keep adding them one at a time, the same way — real evidence first, claims second.
If you want to see the actual queries, screenshots, and dashboards behind any of this, the full report is public. I built it mainly for recruiters and GIS professionals who want to see how I actually work, not just what I say I can do.