From real-time monitoring to forecasting

-

Since 2009, our France database has accumulated over 90,000 items of information on almost 65,000 companies. We track various signals: investments (€1,148 billion), job creation (2.2 million) 👇🏻

...job cuts (1.6 million), fundraising (€162 billion).

Currently, we offer a real-time visualization of past trends, putting us ahead of many comprehensive but delayed or less detailed statistics. Based on our conviction that our data accurately reflects trends in the French economy, and thanks to advances in AI, we are now testing modeling to move from observation to prediction.

Last year, we carried out tests, the results of which will be incorporated into our base this year. We will thus be offering three predictions for the companies in the database:

‣ The dynamics in terms of jobs (positive, negative or neutral);
‣ The number of jobs created or cut;
‣ The number of days until the next news item.
-

The tests show that the reliability of these predictions varies (the most accurate are the riskiest), but the aim is to highlight the remarkable dynamics, positive or negative, in order to classify a company as a dynamic SME, stable large group, etc.

-

(⏪here, the result of models applied to the Pigment company last October. This company has since become a unicorn).

Tests show variable reliability of these predictions (the most accurate are the riskiest), but the aim is to highlight remarkable dynamics, positive or negative, to classify a company as a dynamic SME, a stable large group, etc. (below, the result of models applied to Pigment last October. Since then, this company has become a unicorn).

Other algorithms are also possible, to predict, for example, fund-raising or the evolution of aggregates by sector or territory. These will be developed à la carte or integrated into our database following these first three models.

An exciting new field is opening up for Trendeo and for our customers and future customers!