Skip to main content
Colin Jaffe/4 min read

Mastering API Access with Python's Requests Library

Requests Library Patterns

GET Request

requests.get(url) — fetch data; .json() returns parsed response.

POST with Body

requests.post(url, json=payload) — send data to the API.

Auth Headers

requests.get(url, headers={'Authorization': f'Bearer {token}'}).

Error Handling

response.raise_for_status() throws on 4xx/5xx — handle gracefully.

Master Data Science at Noble Desktop

Noble Desktop's Data Science & AI Certificate covers Python, machine learning, and the modern data science stack.

This lesson is a preview from our Data Science & AI Certificate Online (includes software) and Python Certification Online (includes software & exam). Enroll in a course for detailed lessons, live instructor support, and project-based training.

Accessing the API. Once we've got our URL set up—and this is not a small amount of work—is that learning an API means learning its interface.

Accessing the API. Once we've got our URL set up—and this is not a small amount of work—is that learning an API means learning its interface. What can I hit up to get the data? Once you have that, the Python code to actually access it is very simple. We're going to use the very common requests library.

In fact, this is the one that's so common that it's in the documentation at Alpha Vantage. We're going to say response equals requests.get and give it a URL. You don't have to save the URL into a variable like this, but it's a little more readable this way.

Now, that response from requests.get—this function will hit up this URL, get a response from the server, and put it into whatever variable we save it as. In this case, it's response. And that response will have some things on it.

For example, if we print it out, response has a property called status_code. Let's hit that one up and see what happens. Its code was 200.

All status codes are easily looked up, and it's pretty good to know your status codes. One of the best sites I like for learning status codes is HTTP Status Dogs. We got back 200 OK, which is this adorable dog.

He looks OK. If we had instead created some data on the database by making this request, we would have gotten back a 201 Created code. And there are lots of them, and HTTP Status Codes does a great job showing you all of this information in a very cute, entertaining way.

One that people commonly know is 404 Not Found. You went somewhere on the internet that didn't exist, and the response from the server was “that's not a thing”—404 Not Found.

HTTP Status Codes is great. HTTP Status Dogs is great. It'll tell you, if you click on one of the pictures, exactly what it is.

And if you click here, you'll get the Wikipedia page about all the status codes. And you can read more, of course, in much more detail elsewhere. But for a quick check, HTTP Status Dogs is a great way to familiarize yourself quickly with the basic status codes.

It's also very cute. And you can always access HTTP Status Cats if that's your kind of thing. Now, this status code 200 means everything went well.

If, on the other hand, let's say I have a typo here and run this again, I get this big old error. If I hit up this URL, but let's say I misspell part of it, I now get back an error because I misspelled the domain name.

That's why I got that big error—there was simply no response from any server. This time I contacted the correct server, but I misspelled the interface. This is like clicking on the wrong thing. I misunderstood how to access the interface.

It was like “execute the code block, ” but instead, I clicked over here instead of here, right? I've misunderstood it and, instead of query, maybe I do search. And again, that's not a thing. So I get back a 404.

Again, “Not Found” is the status code. So a common thing to do here is to have some kind of indication from the code that there is some kind of error. We can say if response.status_code!= 200: print("Failed to retrieve data.")

That's fine—it's very general, but there's one too many equal signs. All right, so if I run this, I get “Failed to retrieve data.” A little better than just a number.

Okay, but let's fix that. Instead of search, let's make it query. And now we don't get any printing—no “Failed to retrieve data” message at all.

Next, we'll look at what happens when we do get a 200 OK, and what we can do to actually get our data.