Data Analyst
Golden Eyes Garments
- Location Tirupur, Tamil Nadu, India
- Duration July 2022 – July 2024
- Type Full-time
Golden Eyes Garments is a vertically integrated knitwear manufacturer and exporter based in Tirupur, the knitwear hub of South India. The operation spans design, yarn, fabric, and finished garments across men's, women's, children's, and sustainable product lines, serving international buyers under tight delivery windows. When I joined, the data side of the business was still largely manual. Production tracking lived in scattered spreadsheets, inventory was managed by intuition and end-of-month counts, and weekly reports for the management team were stitched together by hand from multiple sources. The operation was running, but the visibility was reactive. By the time a problem showed up in a report, it had already cost time, fabric, or a shipment window.
There was no system. Production counts were written on paper at each stage of the floor, then transferred into Excel sheets by supervisors at the end of each shift. Inventory lived in stock registers updated by hand. Order data sat with the merchandising team in a separate set of files. Nothing was connected, and nothing was queryable. Pulling a single answer often meant chasing three or four people across the operation.
I spent the first few months on the floor before touching the data. Talking to production supervisors, sitting with the merchandising team, watching how orders moved from cutting to stitching to finishing to packing. That context shaped everything that came after. The most expensive problems were never visible in a single spreadsheet. They lived in the handoffs between stages, in the gap between what the buyer ordered and what inventory could actually support, and in the time the management team spent compiling reports instead of acting on them.
The first thing I built was the foundation. Standardized data collection templates for the floor and the merchandising team, so every shift, every order, and every stock movement was being recorded in a consistent format. I consolidated the existing scattered sheets into a single working dataset, and used Python to handle the cleanup and transformation work that would have taken weeks by hand. None of this was glamorous, but it was the unlock. Once the data could be trusted, every project after it became possible.
The first analytical project tackled production output. I built a structured tracking system across each stage of the manufacturing process and started analyzing where time was being lost. Using Python to join and analyze production logs across stages, I identified bottlenecks that were not obvious from the floor alone, particularly in the handoff between stitching and checking. Working with production supervisors, we adjusted scheduling and line balancing based on what the numbers showed, and output improved by 12% over the following months.
The second project was inventory. Stock overages on yarn and fabric were tying up working capital and creating waste at the end of every season. I built a demand forecasting model that combined historical order patterns, seasonality, and the current order book, using Python for the underlying analysis and Excel as the interface the merchandising team could actually work with. Once they trusted the model, excess inventory dropped by 18%.
The third project was reporting. Every week, hours were being spent manually pulling numbers from different files to produce management reports. I rebuilt the reporting layer in Tableau and Excel, pulling directly from the centralized dataset so updates flowed through automatically. The team stopped building reports and started reading them. Manual reporting time dropped by roughly 15 hours a week, a 40% reduction, and the management team finally had a single, trusted view of the operation.