# Bibliographic Data and Rankings

*This chapter is based on the notes [2022-05-24_SPSE_Rankings.pdf](../../course_notes/2022-05-24_SPSE_Rankings.pdf)*

## Bibliographic Data

Example of DBLP XML:

[Paper xml](https://dblp.uni-trier.de/rec/conf/tacas/Beyer22.xml):

```xml
<dblp>
  <script/>
  <inproceedings key="conf/tacas/Beyer22" mdate="2022-04-29">
    <author orcid="0000-0003-4832-7662">Dirk Beyer 0001</author>
    <title>Progress on Software Verification: SV-COMP 2022.</title>
    <pages>375-402</pages>
    <year>2022</year>
    <booktitle>TACAS (2)</booktitle>
    <ee type="oa">https://doi.org/10.1007/978-3-030-99527-0_20</ee>
    <crossref>conf/tacas/2022-2</crossref>
    <url>db/conf/tacas/tacas2022-2.html#Beyer22</url>
  </inproceedings>
</dblp>
```

Crossrefs:

- ORCID `0000-0003-4832-7662`
    - [lookup on DBLP](https://dblp.uni-trier.de/lookup/)
    - [lookup at orcid.org](https://orcid.org/0000-0003-4832-7662)
- Conference [`conf/tacas/2022-2`](https://dblp.uni-trier.de/db/conf/tacas/tacas2022-2.html)


## Measurements and Rankings

**Be careful about the interpretation of measurements!**

[![](4-Bibliographies-Rankings/measures.png)](4-Bibliographies-Rankings/measures.pdf)


Operational computation of h-index:

1. Sort all papers of authors in decreasing order of citation count.
2. Count downwards as long as citation count >= counter.

Example:

* Publication A: 5 citations
* Publication B: 10 citations
* Publication C: 3 citations
* Publication D: 1 citation
* Publication E: 7 citations

1. Sorted:
    1. B 10
    2. E 7
    3. A 5
    4. C 3
    5. D 1
2. Count until 3 => h=3

![](4-Bibliographies-Rankings/h-index-relation.png)


Common understanding of h-index for a person:
* \> 10: W2 professor
* \> 20: W3 professor
* \> 30: excellent reputation
* \> 40: outstanding

**Heavily biased**:

- age of scientist/journal/institution
- area of research (some cite more broadly, others more narrowly)
- collaboration culture
- size and volume of institutions/conferences/journals
- increasing number of publications per year ([reference](https://ncses.nsf.gov/pubs/nsb20206/publication-output-by-region-country-or-economy))

Other measures:
* h5 for publication venues (h-index for publications of last 5 years)

* i10 for people (#papers >= 10 citations )


## Why Rankings?

![](4-Bibliographies-Rankings/unknown-image.png)

* Milestones / career steps
* Quantify objects
* Select universities (by rank, research fields, ...)
* Funding decisions
* Hiring decisions (select University/Department/Professor for your PhD)
* Legitimization of decisions
* Select publication venues


## Important Rankings and Platforms

* Google Scholar: Scientists (h, i10, #citations, #papers) and Venues (h5)
* [Core Ranking](http://portal.core.edu.au/conf-ranks/) (Australia): Conference rankings by tier (A*, A, B, C, ...)
* [csrankings.org](http://csrankings.org/)
* [csmetrics.org](http://csmetrics.org/)