Revise burst width outliers

Some events appear to have exceptionally wide burst widths. Most of these appear to be associated with `Pushchino LPA`, so these events have to be double-checked and revised according to the relevant publications.

In order to ensure maximum database reliability for population studies, if such observation data fail to appear reliable, all corresponding rows shall be dropped from [`catalogue.csv`](https://github.com/HeRTA/FRBSTATS/blob/main/catalogue.csv).

## Checked Pushchino events
- [ ] [FRB 20160920A](https://www.wis-tns.org/object/20160920a)
- [ ] [FRB 20170606A](https://www.wis-tns.org/object/20170606a)
- [ ] [FRB 20151018A](https://www.wis-tns.org/object/20151018a)
- [ ] [FRB 20180321A](https://www.wis-tns.org/object/20180321a)
- [ ] [FRB 20160206A](https://www.wis-tns.org/object/20160206a)
- [ ] [FRB 20141216A](https://www.wis-tns.org/object/20141216a)
- [ ] [FRB 20161202A](https://www.wis-tns.org/object/20161202a)
- [ ] [FRB 20131030A](https://www.wis-tns.org/object/20131030a)
- [ ] [FRB 20140212A](https://www.wis-tns.org/object/20140212a)
- [ ] [FRB 20121029A](https://www.wis-tns.org/object/20121029a)