Rocket launch API data

There are few publicly available datasets with detailed rocket launch information. Two that I found are LL2 (free) and RocketLaunch.Live (USD 3/monthly). I thought it'll be an interesting exercise to export launch data from an API to see what insights and charts could be created.

I went with the RocketLaunch.Live API simply because I discovered it first while trying to find launch datasets. The API has 7 entities:

• Companies
• Launches
• Locations
• Missions
• Pads
• Tags
• Vehicles

Connecting to the API using a Python script is fairly straightforward:

def get_response(self,url):
	for q in range(10):
		try:
			r = requests.get(url,timeout=10) # Sending GET request to the endpoint/url with a timeout of 10 seconds
            js_response = r.json() # retrieve JSON from the response for further parsing
		return js_response
		except:
			pass

Parsing the JSON response is a little more tricky, as the API only returns 25 elements per call. The response is also nested, with the first portion of the payload containing metadata that we don't need:

"valid_auth": true,
    "count": 25,
    "limit": 25,
    "total": 4483,
    "last_page": 180,
    "result": [
        {
            "id": 3477,
            "cospar_id": "",
            "sort_date": "-287326017",
            "name": "Tiros 2",
            "provider": {
                "id": 12,
                "name": "U.S. Air Force",
                "slug": "u-s-air-force"
            },
            "vehicle": {
                "id": 137,
                "name": "Thor-Delta",
                "company_id": 147,
                "slug": "thor-delta"
            },
            "pad": {
                "id": 90,
                "name": "SLC-17A",
                "location": {
                    "id": 62,
                    "name": "Cape Canaveral SFS",
                    "state": "FL",
                    "statename": "Florida",
                    "country": "United States",
                    "slug": "cape-canaveral-sfs"
                }
            },
            "missions": [
                {
                    "id": 5310,
                    "name": "Tiros 2",
                    "description": null
                }
            ],
            "mission_description": null,
            "launch_description": "A U.S. Air Force Thor-Delta rocket launched the Tiros 2 mission on Wednesday, November 23, 1960 at 11:13 AM (UTC).",
            "win_open": null,
            "t0": "1960-11-23T11:13Z",
            "win_close": null,
            "est_date": {
                "month": null,
                "day": null,
                "year": null,
                "quarter": null
            },
            "date_str": "Nov 23 1960",
            "tags": [],
            "slug": "tiros-2",
            "weather_summary": null,
            "weather_temp": null,
            "weather_condition": null,
            "weather_wind_mph": null,
            "weather_icon": null,
            "weather_updated": null,
            "quicktext": "Thor-Delta - Tiros 2 - Nov 23 1960 (estimated) - https://rocketlaunch.live/launch/tiros-2 for info/stream",
            "media": [],
            "result": 1,
            "suborbital": false,
            "modified": "2023-04-02T22:29:36+00:00"
        }

First we use a python library that can flatten a JSON:

from flatten_json import flatten
flat_json = flatten(js_response) # Using flatten function imported from flatten_json to flatten JSON

We can then use a custom function to take in a flattened JSON and generate a dict for each flattened row's variables (what will become our CSV columns). We then iterate through the dict and write to a CSV file:

def  convert_to_rows(self,flat_json):
n = 0
dct = {}
for key in flat_json: # Reading the key from flattened JSON
	if 'result_' in key: 
		key_name = key.split('_',2)[-1] # Pull the key name from the key and remove the object number from the key, result[120].spacecraft.name becomes spacecraft.name
		if  f'result_{n}' in key: 
			if n not in dct: # Putting the data into the dictionary result item wise
				dct[n] = {key_name:flat_json[key]}
			else:
				dct[n][key_name] =  flat_json[key]
		else:
			n+=1
			dct[n] = {key_name:flat_json[key]}

for row in  dct: # Going through each item in dct
	sub_row = []
	for col in self.columns: # Pulling info for each columns
		if col in dct[row]:
			sub_row.append(dct[row][col]) # Adding to row
		else:
			sub_row.append('NA') # If a column doesn't exist write NA

	self.rows.append(sub_row) # Append the row to self.rows for further processing

def write_to_csv(self,csv_name):

This gives us our full dataset for all rocket launches since 1930:

To visualize the data, we load our CSV into Datawrapper and first create a table for EDA (you can try filtering to only SpaceX launches by searching for 'SpaceX' below):

To validate that our data is reliable, we can compare RocketLaunch's 2024 launch numbers with secondary sources. On average Space Stats reports a -5% discrepancy, while NASA's discrepancy on average is -7%. Overall, the numbers are directionally aligned with minor absolute differences, so I'm comfortable using this data to describe general launch trends.

Finally, we can visualize the API data to look at overall trends (chart is interactive):

Areas for improvement are to automate the data export to Google Sheets and use Datawrapper's Sheets import feature. This will keep the data fresh and updated daily on a recurring schedule.

Stack: VSCode, Python, Excel, Datawrapper
Code: GitHub