.svg)
About the team
“We have a small team but we’re at the heart of many topical issues. We carry out studies for Management, the Marketing department, as well as journalists: we receive quite a lot of requests, but carrying out all these studies takes time, especially when we can’t centralize them anywhere.” — Pierre Mbock, Data Analyst
Le Figaro is a renown French newspaper, addressing ongoing topical issues daily. Le Figaro data team is part of the Subscription Marketing department, but also focuses on the content published on Le Figaro website.
The data team is divided into two groups:
- 2 Business Analysts, who support Business Users with day-to-day request mainly using dashboards solutions ;
- 3 Data Analysts/Engineers, who take care of more advance ad-hoc analyses and data preparation tasks.
Among the various analyses conducted by the team, their main use cases consist of:
- analyzing customer behavior on the app and website (read pages, sessions)
- conducting product-oriented studies, especially concerning subscriptions (measuring sales, conversion rates, usage rates, etc.)
- working with semantic data to determine current topical issues and help writers and publishers choose their subjects
- checking the data source reliability in order to ensure data quality
Le Figaro stack
Le Figaro data team uses different tools throughout their stack:
- One data source: all their data is stored in BigQuery
- dbt enrichment: it enables them to generate automatically descriptions, sources, tests, model SQL, etc.
- Various dashboards solutions: general studies are constructed on Looker Studio and Tableau
- A notebook solution: advanced analytics such as ad-hoc reports are built directly on Husprey
“Husprey fits perfectly in our technical stack. It is well-suited for small data teams who need to pass a milestone in terms of productivity. We’ve noticed a real difference after we started using it.” — Quentin Barigault, Lead Data Analyst
The Figaro challenges
As a media provider, new data is continuously generated, and Analysts need to assess the relevancy of each topic. Agility is the number one priority.
To begin with, Le Figaro needed to improve the execution time of each analysis. The quicker the feedback loop is, the more impact on decision-making. As a matter of fact, publishers can’t afford to wait 2 weeks before knowing whether a topical issue is still topical or not.
The second challenge was finding a way to gather knowledge, as it was scattered in multiple tools, such as on-off local Jupyter notebooks, non retrievable BigQuery sql queries and GSlides. The copy-pasting, the maintenance costs and the time spent looking for previous content were all overwhelming.
Last but not least, fostering collaboration was also a big requirement to the team. Le Figaro believes that everyone in the company, whether they are data literate or not, should be informed of the analyses and insights. Choosing a tool that could also improve data literacy among the stakeholders was a big win.
Husprey: agility and knowledge
Husprey notebooks are designed to deal with specific analyses that do not have use in dashboards, such as one-off analyses. Le Figaro Analysts can quickly query the data available in BigQuery and provide reports on a dedicated topic on the spot.
As a notebook is fully shapable, analysts can start and work through their analysis as they wish. This is also true for the visualization used. Analysts can easily switch from a percentage table to a bar chart and therefore give their insight a totally different look, taking into consideration the future readers of the report.
“We used to use Python: it worked very well, but it was a lengthy process. And most importantly, we couldn’t share Python language to non-data teams – it would have been rather rude – so we had to take a screenshot and send it by email…We used too many tools.” — Quentin Barigault, Lead Data Analyst
This resonates with the agility found in the data storytelling of reports, which asses the overall impact of the analysis. Husprey notebooks are fit to suit any stakeholders.
Analyst are in charge to build the report to convey the right message and takeaways. For instance, they can set the overall tone of the analysis — as positive, negative or neutral —, add quotes, or even references to previous analyses. Without these features, sharing the analysis internally with other teams would be quite difficult, as they would first need to extract the data, analyze it and then paste results in GSlides.
To foster collaboration, the team uses the comment feature to pinpoint and iterate on a topic — it's a team favorite. Each study isn’t fixed, meaning it can evolve depending on the stakeholders’ needs.
“What we want is to make the study more accessible by simplifying it, so that anyone can read it and gather the same insights as us even if they have never used SQL. If we export this study as PDF, there is no need to rework it: we can share it right away.” — Pierre Mbock, Data Analyst
Finally, Husprey helps the team save considerable time building analyses. Queries and analyses are searchable and easily reusable, analysts can now re-launch a study that was previously done by a colleague 6 month ago. Combined with automatization – whether it be refresh or automatic launch of the notebook – this reproducibility allows Analysts to implement recurring analysis more easily.
All in all, Husprey has made a significant impact on Le Figaro data team daily operations. More time is focused on actual analyses and reflection, and clear and actionable reports are provided to Business Users.
“When we took part of the Husprey adventure, we were looking for a tool but we didn’t really know what to expect. After one year, I strongly recommend this tool – whether it be for improving productivity or centralizing and formalizing studies.” — Pierre Mbock, Data Analyst
