app id and reviews

[PythonCoderUnicorn]

Hi appler,

I have been enjoying testing your package, thanks for making it. Here are some of my issues with the package.

1. In your documentation you do not explain how the user finds or gets the id which is required for your functions. Only later after checking the HTML tags did i realize the id is in the url.
2. using the search_apple() function first then that provides the artist id for the lookup_apple()

I find that using the functions as below to be more clear that both iTunes and Apple App is being searched

song = "anti-hero"

apple_itunes = search_apple( song, country="ca")   
  
# which returns Taylor Swift and shows `artistId` 
# apple store lookup

artist_id = 159260351

lookup_apple(id= artist_id, country = "ca", sort = 'recent')

3. in the documentation for get_apple_chart_postion() you have a long number for an id for some app called Apollo.
   a) I find this name very confusing as the name Apollo is not specific enough for the app by name, I tried finding the specific app, but i found Apollo for Reddit among other apps. Can you please be more specific ? After some time, i finally understood, by using the website and using the name and id of this app did i confirm that this app is Apollo for Reddit and is searching the UK for data. https://apps.apple.com/ca/app/apollo/id979274575

b) Using the MacOS desktop app i have no way of getting these specific ids, only by searching on the website.
c) perhaps there should be a defined list of app ids ?

4. for the get_apple_review() again by going to the website can i get the id. https://apps.apple.com/ca/app/github/id1477376905 . I noticed that the data i get back is max 500 like it says in the documentation but i am finding that on the website data does not match my returned data. The reviews from the website are older than the very recent reviews.

• https://apps.apple.com/ca/app/github/id1477376905?see-all=reviews

but my code is

github = 1477376905
app_reviews = get_apple_reviews( github, "ca", all_results = T, page_no = 1)

which has a max of "2022-11-20 20:17:06 UTC" and a min of "2020-03-17 18:41:28 UTC", is this a limitation from Apple that only most recent reviews are available?

---

Anyway, thanks for this package. 😸

[ashbaldry]

Hi @PythonCoderUnicorn, thanks for the feedback! Clear documentation isn't my strong point so anything to improve it is a plus 😄

1. There is definitely a lack of information about how to get the ID of an application. I like how you've explained it in point 2., and there should be information about how to extract the ID from a URL.
2. Good point, as well as the example you mentioned for lookup_apple, an example of an application should be included.
3. I should use the same method in 2. by including the lookup_apple first to make it clearer.
   a. I'll change the example to the GitHub app - it is more salient to data science members than Apollo
   b. Didn't know that about MacOS - good to know!
   c. Unfortunately there are currently around 2 million apps, and new ones constantly being added, so don't think I can store that information easily, unless you were thinking of only having a reduced list of well-known applications (plus country availability)?
4. Interesting that you can see earlier reviews on the site, the earliest I can see is the one in the query (it also only shows me 10 reviews even though it says 'see-all=reviews' in the URL). Apple's documentation on this RSS is non-existent, so had to follow a couple of SO questions. That is where I got the 500 max reviews, and it is sorted by most recent first.

[PythonCoderUnicorn]

As for 3-c,

I understand there are way too many apps to list and store here and load into the package, so i was thinking of a predefined list of apps. I could make a issue or PR with a list for Canada so that way there is a vector of ids that could be ready to go for use. If user wants a different app id then show how to find it or something.

As for 4,

I was trying to compare what I am seeing on the website and what I can search in the data. I tried GitHub app, the Apollo for Reddit and Twitter, all three showed older reviews but the data returned very recent ones. As for Apple documentation, I understand it being a pain, so the fact that this app runs nicely is a great achievement !

The main goal of this function is achieved and works, the documentation perhaps could mention that the data fetched is very current or may not be comparable from the website. The data from the reviews is another application for NLP which I enjoy so that is fun to have another tool in the R skillset.

---

thanks again 👍

[ashbaldry]

If you are happy to do it for Canada, that would be appreciated! From that I can judge what to include for UK, then see which common apps are available around the rest of the world.

I think I have found the reason for the older reviews too! Apple has 4 sorting systems: most recent, most critical, most helpful and most favourable. I have hardcoded "most recent" in the RSS pull however it shouldn't be hard to include the other sorting options.

I'll keep in mind everything you've suggested and am going to update the function documentation and update the README to use a vignette instead. If you'd like to review the changes, I can let you know when I'm ready to merge.

[PythonCoderUnicorn]

Okay, I will start making a list or a dataframe of apps and ids for Canada.

I did change the sorting of the reviews on the website to see if any of the data matched but it did not.

Yes, I would like to be involved, I like this package and want to help in any way. I also make hex logos as well if that is something that you want.