two small suggestions

[yonggui-yan]

Thanks for sharing this useful tool—I really appreciate it.

The interactive map in HTML is great for review. However, for my application, I took a screenshot. I’m wondering if there’s a way to automatically save the map in other formats, such as an image file, when I run the original code.

Most of the information is accurate, though I did notice some discrepancies, possibly due to affiliations having multiple locations. I’m curious whether the map is generated from the data in the "citation_info.csv" file. Additionally, would it be possible for users to manually adjust the citation_info.csv file before generating the map? If so, it would be helpful if the user guide could include instructions for this process.

Thanks again.

[ChenLiu-1996]

Hi Yonggui,

These are very good suggestions. I will try to incorporate them when I have time.

To answer your question: "I’m curious whether the map is generated from the data in the "citation_info.csv" file.", the answer is that currently it's not saving and loading the data but rather just saving the intermediate results. However, since you proposed a possibly useful feature, I can change the logic for accessibility.

Best, Chen

[scienide09]

I just recently found your project. I think the suggestions in this thread are both good. I too would like to see if I could find and update some missing affiliations and add that data to the map.

[ChenLiu-1996]

Update on these suggestions.

1. For manual editing of csv, I have updated the code to support that. See version 4.5 and the changelog (https://github.com/ChenLiu-1996/CitationMap?tab=readme-ov-file#changelog).
2. For image file output, I have tried for a bit without finding a good solution. Also, since this is a less critical issue I will ignore it for now. You can get most of the functionality and convenience by zooming in/out on the HTML and doing a screenshot.

[scienide09]

I really appreciate your willingness to adapt this. Now that I've finally had a chance to look at the 4.5 changes, I want to be sure I'm understanding right.

I understand that when scraping, the message "[WARNING!] Could not find author links for [paper title]" generates when there is no Google Scholar profile attached to that citing work, and therefore the affiliation is not known. These entries are scrubbed before generating citation_info.csv, such that the total number of citing affiliations recorded is lower than the action number affiliations.

Instead of scrubbing those entries, IMO what would be useful is if the instances where the affiliation is unknown, these are still populated into the citation_info.csv file (with just citing author name, citing paper title, and cited paper title data). This would allow us to manually edit the .csv to add the missing affiliation, geo coordinates, etc., based on what we do have, then generate the map again using 'parse_csv=True'.

[ChenLiu-1996]

That is a very good point. Really busy in the following two days but will try to fix it at some point.

[ChenLiu-1996]

I think version 4.6 addresses this issue. Please ask me to reopen if it does not help.