Lyon is France's 2nd tech hub. Your data analysis is still stuck in manual mode.

Lyon has 60,000 digital jobs, the H7 incubator, top-tier programs at Emlyon and INSA. The tech ecosystem is strong. Yet most Lyon businesses, including those that call themselves data-driven, still run on copy-pasted Excel dashboards every Monday morning. That's reporting, not analysis.

Take pharma. Sanofi Pasteur in Marcy-l'Etoile, BioMerieux, Boehringer Ingelheim: these companies generate massive volumes of production, quality control, and clinical trial data. But the 80-person pharma subcontractor in Gerland? They track quality in a shared spreadsheet on OneDrive. Nobody digs into the correlations between production batches and non-conformity rates.

Same story in manufacturing. Renault Trucks in Venissieux, the Vallee de la Chimie chemical corridor, mid-market manufacturers across the Rhone valley: maintenance, logistics, and production data sitting in ERPs. The plant director knows there's value in there. He just doesn't know where to start.

Lyon's food industry, as you'd expect from France's gastronomic capital, generates its own data too. Danone nearby, dozens of SMEs in food processing. Sales, inventory, seasonality data exists. Nobody asks it the right questions with the right statistical methods.

What I do for businesses in Lyon

Exploratory analysis: finding what you weren't looking for

You have a business question? I take your raw data and dig. Principal Component Analysis (PCA) to reveal hidden structures in your production or quality data. Correlation circles to understand which variables move together. Clustering to segment your customers, products, or markets. For a manufacturer in Venissieux, that might reveal correlations between machine parameters and production yield. For a pharma lab in Gerland, patterns in quality control data that nobody had spotted. This isn't reporting. It's exploration.

Ad-hoc analysis: concrete answers to your questions

"Why have our maintenance costs spiked in the last six months?" "Which customer profile generates the most margin across our distribution network?" "Is our pricing consistent between our Lyon, Grenoble, and Saint-Etienne offices?" Every business question has an answer sitting in your data. I build the analysis end to end: extraction, cleaning, statistical exploration, results visualization. You walk away with a clear deliverable and actionable recommendations. Not an 80-page report.

Opportunity detection: making your data work for growth

Most mid-market companies in Lyon use their data to look backward. How much did we sell, how much did we spend, what was the conversion rate. I do the opposite: I use your data to look forward. Which market segments are underexploited across the Lyon-Saint-Etienne-Grenoble triangle? Which products could be bundled based on actual buying behavior? Which correlations between your marketing actions and your sales have never been measured? That's the kind of analysis that turns a data cost into an investment.

How it works

1. Scoping call (free, 30 min)

You tell me your business question. No need for a structured brief: "we feel like we're missing something in our customer data" is a perfectly fine starting point. I tell you if the analysis is feasible and what I need to get started.

2. Data access and initial exploration

I get your data (export, API access, SQL database, whatever the format). I do a first pass: data quality, usable volume, early patterns. Within a few days, you get an honest assessment of what the data can and can't tell you.

3. Deep analysis

This is the core of the work. PCA, clustering, statistical tests, correlation models: I apply the methods that fit your question. Not data science for the sake of it. Every technique is chosen because it solves a concrete problem.

4. Deliverable and recommendations

You get a visual report with the key findings, the visualizations that matter, and concrete recommendations. I walk you through the results on a call, answer your questions, and make sure you know exactly what to do with what we found.

Tech stack

Pythonpandasscikit-learnmatplotlibseabornSQLBigQueryJupyterLooker StudioTableau

Frequently asked questions

You're based in Tbilisi. How does that work for a client in Lyon?

I grew up in Haute-Savoie, an hour from Lyon. I know the local business fabric well. Today I work remotely from Tbilisi: video calls, shared reports, async updates. It works better than most on-site setups. For kickoff or sensitive phases, I come on-site. Lyon is a direct flight away. But your mid-market company in Villeurbanne doesn't need me sitting in their office. They need their analysis delivered on time.

How much does a data analysis cost?

Depends on the complexity of the question and the data volume. A focused ad-hoc analysis is a few days of work. A full dataset exploration with segmentation and modeling, more like one to two weeks. We scope the budget together on the first call, and if the project doesn't justify the investment, I'll tell you.

Why not just go with CGI Lyon, Capgemini, or Atos?

Lyon's big consulting firms are great for large-scale transformation programs. But if you need one person who digs through your data for two weeks and delivers clear answers, that's not their model. At those firms, a salesperson sells the project, a project manager scopes it, a junior consultant executes it. With me, one person who does the work himself. I work with companies that are too small for Capgemini and too ambitious for "my nephew knows Excel."

We don't have a data scientist in-house. Is that a problem?

That's exactly why I'm here. You don't need a permanent data team to ask a one-off question to your data. I run the analysis, present the results in a format your team understands, and document everything so you can reuse the insights.

Is our company mature enough for data analysis?

If you have data (Excel files, a CRM, an ERP, logs), there's something to work with. Perfect data maturity doesn't exist, and waiting to reach it before starting is a classic mistake. The initial exploratory pass is precisely how I assess what your data can tell you. If it's too incomplete, I tell you before billing for an analysis that would go nowhere.