Case Study 40-A: Priya Okonkwo — Eighteen Months of Python at Acme Corp

Subject: Priya Okonkwo, Head of Data Analytics, Acme Corp Timeline: 18 months after beginning Python Role at start: Senior Analyst, Sales Operations Role now: Head of Data Analytics (newly created position)

Before Python: The Excel Analyst

Before Priya wrote her first Python script, she was considered one of the best Excel users at Acme Corp — a distinction she earned honestly. She could build a pivot table faster than most people could open the data, she maintained eleven interconnected workbooks that represented years of institutional knowledge, and she had a well-deserved reputation for accurate, reliable reporting.

She was also, by her own account, exhausted.

Every Monday began with the same four hours: opening four regional sales files, pasting them into a master workbook, applying the formulas she had memorized years ago, checking for the inconsistencies that crept in whenever the regional teams changed their column names or date formats, and finally building the chart that Sandra Chen would present to the executive team that afternoon.

The process had been refined over two years. It was reliable. It was also, Priya thought quietly, something a computer should be doing.

She had said this to Marcus Webb, Acme's IT director, once. He had agreed, and gently suggested she might want to look at Python. "It would take you maybe a week to automate that," he said. She had nodded, gone back to her desk, and not thought about it again for several months.

The First Script: A Monday Morning Changed

The moment that changed things was not a dramatic realization. It was a Wednesday afternoon when Priya was re-running a report for the third time because someone had sent her a corrected data file after she had already sent the report. She had made an error in the manual transfer of one of the corrections, and the corrected version she sent contained a new error.

She pulled up Python documentation that afternoon. She did not understand most of it. She found this textbook that evening and started Chapter 1.

By Chapter 11, she had a working version of the Monday report in Python. It was not elegant. It had hardcoded file paths and no error handling. But it ran, it produced the right output, and it ran in 47 seconds instead of four hours.

She did not tell anyone she had built it. She ran it in parallel with her manual process for three weeks, comparing outputs to verify they matched. When she was satisfied they did, she let the script run the report and spent the four hours she recovered working on a customer concentration analysis that had been on her "someday" list for six months.

The Dashboard That Changed Her Career

The customer concentration analysis became a dashboard. The dashboard began as a Jupyter notebook — not yet deployed anywhere, just running on her laptop. But when Sandra asked "Is there any way to look at account concentration by region in real time?" in a Monday morning briefing, Priya said, without entirely planning to: "I actually built something for that. I can have it up this afternoon."

She stayed late and turned the notebook into a basic Flask application running on Acme's internal network. Sandra saw it at 4 PM. By 5 PM she had forwarded the link to the CFO. By the following Monday, Marcus had moved it to a more reliable server and Sandra was asking Priya what else she could build.

The answer, it turned out, was quite a lot.

Over the following eight months, Priya built five tools that entered production use at Acme:

1. The Monday Report — The automated regional margin report that started everything. It has run reliably for forty-one consecutive weeks at the time of writing. Priya has not touched the underlying code in eleven weeks.

2. The Customer Dashboard — The Flask application showing account concentration, tier breakdown, and at-risk accounts. Used by Sandra, the CFO, and four regional directors every Monday morning.

3. The Inventory Alert — A script that monitors stock levels in Acme's inventory database and emails Marcus and the warehouse manager when any SKU falls below reorder point. It prevented three stockout situations in Q3 and paid for itself in the first month.

4. The Sales Forecast — A compound growth model with three scenarios (optimistic, base, conservative) updated monthly. Sandra uses it as the primary basis for her board presentations. The CFO has referred to it twice in board documents.

5. The Churn Predictor — Priya's most technically ambitious project: a logistic regression model trained on 18 months of customer transaction data that scores each active account on churn probability. Sandra considers it the most valuable analytics tool the company has ever had. The model has correctly flagged four accounts that subsequently reduced orders — two of which were retained through proactive outreach that Priya suggested based on the model's feature importance scores.

The Promotion: Head of Data Analytics

In January, nine months after building the first script, Priya was promoted to a newly created role: Head of Data Analytics.

The promotion letter, which Sandra read aloud to the assembled team, cited her "demonstrated ability to build technical infrastructure that directly supports business decision-making" and her "entrepreneurial approach to identifying and solving problems before they are formally assigned."

"It was the nicest thing anyone has ever said about me professionally," Priya said later. "And none of it was true eighteen months earlier."

The new role came with a budget and two direct reports: Kenji Yamada and Fatima Al-Hassan, both recent graduates with strong statistics backgrounds but no Python experience. Priya was tasked with building the analytics function from scratch and training her team.

She used this book as one of her primary teaching resources. She assigned one chapter per week and ran weekly code review sessions where Kenji and Fatima walked her through their exercise solutions. She found, as every teacher finds, that explaining a concept forced her to understand it at a level that learning it passively never had.

The Conference: A Different Kind of Recognition

In October, Priya presented at the Regional Business Intelligence Conference — a two-day event attended by analytics practitioners and business intelligence leaders from companies across the region.

Her talk was titled: "From Spreadsheets to Python: How One Analyst Rebuilt the Analytics Function at a Mid-Sized Distributor."

She was nervous in a way she had not been since college presentations. She had built the slide deck over two weeks, rewritten the opening paragraph four times, and practiced the demo three times the morning of. She knew the material completely — she had lived it — but she had never presented it at this scale.

The talk went well. More than well. The Q&A ran six minutes over the scheduled time. Three other attendees approached her afterward, gave her their cards, and asked follow-up questions about how she had handled the deployment. One was building something similar at a manufacturing company. Another was trying to convince her management to invest in Python training and wanted to cite Acme's experience.

Priya left the conference with a clearer sense of what she had actually built — not just the tools, but the case study. Her story of how a single motivated analyst transformed the analytics infrastructure of a mid-sized company using free tools and a few months of self-directed learning was, she realized, worth telling. She plans to write it up properly.

How She Evaluates BI Tools Now

When Acme's procurement team began evaluating enterprise BI tools — Tableau, Power BI, Looker — in Q4, Priya was included in the evaluation committee.

Her perspective had changed in a way she found useful. She knew exactly what these tools were doing under the surface. She could look at a Tableau dashboard and identify the underlying groupby-aggregate operations. She could look at a Power BI data model and recognize the relationship structure she would have built in SQLAlchemy. She could read a vendor's claim about "flexible data connectivity" and know precisely what that meant in technical terms.

This did not make her dismissive of commercial tools. It made her more useful in evaluating them. She could ask the right questions, understand the answers, and make accurate cost-benefit comparisons — including the "could we build this ourselves and at what cost?" calculation that she was uniquely equipped to make.

The committee ultimately recommended a modest Power BI deployment for the finance team, keeping Python for the analytics team's more complex and custom work. Priya wrote the recommendation memo. It was three pages. It was accurate.

Priya's Advice

When Kenji and Fatima ask her what they should focus on as they build their Python skills, she tells them the same three things:

First: Build something real from the start. Not a tutorial exercise, not a toy dataset — something connected to actual work. The moment it runs and produces a real result is the moment it becomes real to you.

Second: Do not optimize prematurely. Write the simple version first. Make it work. Make it clear. Then, and only if it is actually too slow for actual use, make it fast. Most business tools do not need optimization. They need to be correct and clear.

Third: Document while you build, not after. Your future self, your colleagues, and your users will thank you. The documentation is half the product.

She adds, when she thinks they are ready to hear it:

Fourth: The goal is not to become a software engineer. The goal is to become a business professional who can build things. Those are different goals, and they require different tradeoffs. Know which one you are pursuing.

Snapshot: Priya at Eighteen Months

Role: Head of Data Analytics, Acme Corp Team: 2 junior analysts (Kenji Yamada, Fatima Al-Hassan) Tools in production: 5 active Python applications serving 12+ internal users Conference presentation: 1 (Regional BI Conference, October) Next goal: Building a formal analytics roadmap for Acme's next fiscal year, including evaluation of a cloud data warehouse for the analytics team

What she would tell her eighteen-months-ago self:

"Start immediately. Not when you have more time. Not when you feel ready. Now. The readiness you are waiting for comes from doing the thing, not from preparing to do the thing."