Fix a header and add a link to the dataset page

README.md

@@ -35,6 +35,8 @@ WHERE (date_part('year', uploaded_on) >= '2018') AND skip_reason = ''
 GROUP BY project_name, project_version, project_release
 ```
 
+You can download the csv and the sqlite files [here](https://dataset-dl.liris.cnrs.fr/pypi-projects-backends/)
+
 ### CSV to sqlite3
 
 An [SQLite](https://sqlite.org/index.html) database is created with this
@@ -242,7 +244,7 @@ FROM version_releases_types
 GROUP BY project_name;
 ```
 
-## New analysis
+## New analysis
 
 The first created table, `pyprojects holds 8 742 125 `project_release`
 (source and binary packages) from 410 944 `project_name`.