sK: FIS data import - Githubissues

mslehre / text-embedding

LLMs and embeddings for Uni GW texts

2 stars 0 forks source link

Open MarioStanke opened 1 year ago

MarioStanke commented 1 year ago

Adapt the new data from Forschungsinformationssystem for our existing code and data

sKt1: bring tables from Daniel Fredrich into our format (3 steps) publication lists, list of institutes
sKt2: merge with existing data from WebOfScience (unique IDs, concatenate publication lists)
sKt3: reconstruct 2D plots, option for thinning out point, e.g. by number of publications) (2 steps)

MarioStanke commented 1 year ago

Structure of tables:

table name: column name 1, column name 2, ... the columns are tab-separated

Table inst_abbrev.tbl: long_name, short_name The long_names are unique in this table.

The short name is for example Physik instead of Institut fuer Physik.

Table inst_color.tbl: short_name, color

The color is a consecutive integer. The short_name is unque in this table.