July 2nd meeting
Attending: Loïc Charles, Guillaume Daudin, Béatrice Dedinger, Paul Girard, Guillaume Plique, and Pierre Hollegien.
1-Team Management
Guillaume Plique will be on vacation for 20 days in August and will remain with us until the end of January.
2-Budget
All the budgets are in; Guillaume Daudin must finalize it and prepare the documents for the ANR
3-Ricardo
On Monday 6th and Tuesday 7th, an IT team will work with Paul, Béatrice, Guillaume P. and Guillaume D. on remaining issues on Ricardo visualization platform. This meeting will tackle slight changes (add new options such as “order per name”, “filter”, changing some titles) and the creation of new variables such as the different flows views (estimated, reported, “sum of partners”, world unknown, best guess).
Discussion on persistent issues (sources standardization, view prioritized, etc…) will also take place there.
Moreover Béatrice D. will participate to a meeting in October gathering schoolteachers in order to collect opinions and advices for the visualization platform. Yves Poncelet, academic inspector, will assist to this meeting.
Finally, Ricardo’s budget is set to 15,000€ for the next three years. Accordingly an intern will be recruited for 4 months (4,500€) to develop and design the website. Paul G. has already found a suitable candidate. The internship should take place from September to December.
4-Retranscription
Two new interns will join the team for two weeks each. Anthony Rebours will join us on July 15th and Elisa Maria Tirindelli should join us later as well.
Guillaume D. might also enroll another intern but some issues remain.
Transcription from Bayonne, Rennes, Rouen, Caen and the Colonies must be reviewed by Loïc but are completed.
Thasectra is implementing data from Marseille. Next they should deal with La Rochelle 1789.
Besancon and a part of Lyon will be inputted by Anthony Rebours. Currently Cyril Mouton is working on colonial flows and Elisa Maria Tirindelli should work on data available for 1792.
5-A new tool for visualization: Neo4J
Neo4J is a graph database that must enable us to visualize flows according to their characteristics (import/exports, direction, commodity, etc…). Guillaume P. and Paul G. have been working only for a short while but they manage to make a demonstration with 10,000 flows. In our case we cannot use a relational database since the type of flow (location where it was recorded, year, etc…) can change the characteristics we attribute to one commodity. We thus need to use a network structure.
Remaining problems:
- Products from Belgium customs and the Sund are mixed in our database è we must integrate metadata for each commodity that enable us to know the source.
- How to enable the user to add corrections:
- Integrate a filter combine with an orthographic search
- The system works only with 3 columns : the distribution key that must remain unchanged, the selected column on which we are applying corrections, and finally the column following the selected one (to match with classifications)
Neo4J will be tested for the next week in order to come up with solutions to the remaining issues. We will then draft a final version.
To help with unit standardization, Loïc C. has found an interesting book giving information on measures used in south-west of France.
Next meeting is scheduled at 2 pm, at the Médialab on July 20th. It will be a small-committee one where we will discuss of the various tools required. Another meeting is scheduled on September 10th at 2pm at the médialab.