This repository allows to use the jupyter notebook made in corresp-humboldt-dataviz in a browser. The user no longer needs to clone the whole repo and launch the jupyter notebook in his terminal in order to use it. This repository offers better accessibility than the corresp-humboldt-dataviz.
Access to jupyter notebooks and map visualisations: corresphumboldtlite
In order for the map visualisations to be presented in the jupyter lab, the code had to be adapted. Furthermore, these are not entirely the same functions as the one in corresp-humboldt-dataviz. Also, within this repository only the jupyter notebook presenting the data visualisations is available, not the dynamic search within the data (search.ipynb in corresp-humboldt-dataviz).
This experimental project seeks to discover, explore and visualize the correspondence of Alexander von Humboldt. The idea started with the catalog of Alexander von Humboldt's letters held at the Berlin-Brandenburg Academy of Sciences (BBAW). Up to now, the collection is only accessible via an index card system established in the 1950s. Only the research aid has been digitally reproduced.
The original idea was to make at least a part of the collection digitally accessible and to discover it with the help of new research tools. With this goal in mind, the catalog was to be correlated with modern manuscript databases.
The used dataset come from 3 different sources:
- The Kalliope Verbund: data of the letters sent and received by Alexander von Humboldt (AvH) have been retrieved from the Kalliope's API in Dublin Core format.
- Bibliothèque nationale de France (BnF): data were retrieved in csv format on the user-friendly API
- American Philosophical Society : data were retrieved in EAD format.
The three dataset were cleaned and homogenised in order to be able to query them. To visualise the letters on a map, geopoint, geoname ID and humboldt digital edition identifier (edh id) were also added.
In the digital collection there are 3465 letters written by Alexander von Humboldt and 1467 letters sended to him.
Several libraries were used for data visualisations, among them the main ones are: ipywidgets, ipyleaflet, matplotlib, pandas, numpy.
corresp-humboldt-dataviz is developed by Axelle Lecroq, intern in the BBAW's project "Travelling Humboldt - Science on the Move". Corresphumboldtlite enhances corresp-humboldt-dataviz in order to give an easier access to the jupyter notebook and to use them in browser.