Rerunning all the analysis notebooks for the report (#81)

* Some edits to the naming of level b and c groups in build taxonomy, add config for new analysis

* Changes to locations_to_nuts to allow for new data structure

* Allow for tfidf params in naming of levels

* Rerunning all analysis and applications notebooks

* Replace report images with updated ones

* Add new images to main readme

* Add image of 2022 skill clusters

* Move get duplicate sent ids to pipeline and reformat output to be more memory efficient

* Save out duplicates only

* Darken green in examples jpg

* Update analysis notebooks to include the additional data from the duplicated sentences

* Add combination of duplicated sentences dictionary and save, and describe script in the readme

* Delete some lines that shouldnt be there from cluster embeddings

README.md

@@ -8,7 +8,7 @@ This repo contains the source code for this project. To read the full technical
 
 An overview of the methodology, coloured by the three main steps to the pipeline, can be visualised below:
 
-![](./outputs/reports/figures/methodology_overview_pipeline.jpg)
+![](./outputs/reports/figures/Jan%202022/methodology_overview_pipeline.jpg)
 
 ### Pipeline steps
 
@@ -23,10 +23,10 @@ More details of the steps included in this project, and running instructions, ca
 
 This repository also contains various pieces of analysis of the taxonomy. These are discussed in the main analysis [README file](skills_taxonomy_v2/analysis/README.md).
 
-<img src="./outputs/reports/figures/hierarchy_numbers.jpg" width="700">
+<img src="./outputs/reports/figures/Jan 2022/hierarchy_numbers.jpg" width="700">
 
 #### Examples of the hierarchy
-<img src="./outputs/reports/figures/taxonomy_example.jpg" width="700">
+<img src="./outputs/reports/figures/Jan 2022/taxonomy_example.jpg" width="700">
 
 ## Setup