Last Updated on January 8, 2022
- What are Data Collection Jobs? Why will I get paid for my data?
- Are data collection jobs safe? How will my data be used?
- What type of content can I submit, and how much can I earn?
- The issue with data collection jobs (and how to avoid them)
- How to avoid getting your submission rejected:
- Data Collection Websites Megalist (Updated for 2022)
Data Collection jobs are one of the most fun and lucrative ways to earn money online.
Modern AI algorithms need a lot of raw data for training. The more data you feed the algorithm, the better it becomes.
Hence, there are a lot of companies that are willing to pay you for your photos, videos, audio recordings, and whatnot.
What are Data Collection Jobs? Why will I get paid for my data?
I get it. Sharing your personal photos and videos for the development of software does not sound like a compelling argument.
But guess what? If you are using Google Photos, you are already sharing all your images/videos with Google, for free! (read their T&C if you don’t believe me).
They use them to improve upon their image and video-related algorithms (the reason behind Google Pixel’s amazing camera performance) and for various other internal projects.
Update: Google Photos is just an example. There are various other online services that you use, where you can lose custody of your data. Such as Instagram!
According to Instagram, once you upload any photo to it, it becomes their legal property. You lose rights to it.
So stop giving away your data for free, and instead get paid for it.
Even if you don’t use Google Photos or Instagram, just go through your Camera Roll and look through your collection.
How often do you actually look at your photos? Are they really private enough? What’s the last time you actually opened them?
Wouldn’t it be great if you can get paid for them?
Are data collection jobs safe? How will my data be used?
They specifically request you to protect your personal information, and also explain the purpose of the data collection job (mostly training ML/AI algorithms).
Most of the data collection jobs you encounter will not ask for any of your personal or identifiable information.
On the rare occasion that they do ask for your personal info (such as an identity card/bank card for “research purposes”), you should have the common sense of skipping them.
I have only come across one task so far that felt shady. It asked for my govt. issued ID card, convincing me that it will be used for training some face recognition software. I simply skipped that task, as it felt too personal for me.
Make sure you never give away any personal or sensitive information for doing a data collection task, no matter what they say.
So there is a fine line here. You need to have some sense as to what information you give away. Most legit jobs do not ask for your sensitive information. Be safe out there!
What type of content can I submit, and how much can I earn?
I have participated in a tonne of data collection projects, and have submitted everything from photos of street signs, videos of my hands performing various gestures, and audio recordings of giving various commands to voice assistants.
There are data collection projects that are incentivizing you for uploading pictures of your pet, your child, textbooks, receipts, household objects, etc.
Earnings vary from project to project, but as a general rule of the hand, you get 0.10$-0.50$ per image, 1$-5$ per video, and 0.05$ – 0.20$ per audio recordings. The price varies depending on the complexity of the task.
Every month, I earn around 20$-100$ extra by just doing some data collection projects on the side.
An interesting thing to note that several of these projects repeat across the different platforms (Appen, Clickworker, Neevo, etc.) so you can actually earn multiple times from the same data set.
I have earned over 50$ by submitting photos of Textbooks, 20$ for submitting videos where I pretend to have a video call under different lighting conditions, and around 40$ recording voice commands in various regional languages.
There’s no limit on the earning, and the best part is that it takes only several minutes to complete the tasks. You are actually earning anywhere between 10-20$/hour if you calculate it that way.
The issue with data collection jobs (and how to avoid them)
One problem that turns people away from data collection jobs is when their data is rejected. This is quite a common occurrence with people.
Earlier I used to believe that it happens when people don’t follow the guidelines (how to take photos, rules, etc.) correctly.
But turns out that the companies themselves are notorious for doing this. I am usually paid only 70% of the promised amount for doing most data collection jobs.
Don’t know how but most of them simply reject a part of my submission (for vague reasons).
Sites like Lionbridge/TELUS, Clickworker, Oneforma, etc. don’t do it as much as sites like Neevo. The reason behind this is that the former sites have internal teams to check your submissions, while the latter ones verify your submissions through other crowd workers.
It’s still worth giving sites like Neevo a shot because they have some of the best pay rates in terms of data collection projects. Appen has the worst rates.
Update: Oneforma has had a slew of Data Collection Jobs recently. And their pay is very competitive, with some jobs paying as much as 75$ per set.
How to avoid getting your submission rejected:
- Read the guidelines very carefully. I cannot emphasize it enough. There are very minute instructions for you in the guidelines that if missed, can get your submission rejected.
- There are instructions on how to frame the image/video, the kind of lighting and background conditions required, silent/noisy environment (for audio recordings), etc. that should be thoroughly read and followed while creating your dataset.
- No images from the Internet. I don’t know why a lot of people are hell bent on cheating the system. And then they complain about their submission getting rejected. If they wanted images that are freely available on the Internet, they wouldn’t be paying you for it.
- Whatsapp images are not allowed. This is because Whatsapp heavily compresses the content you share (Photos shared on whatsapp are less than 2 Megapixels in size). Most of the data collection projects require larger and higher quality photos.
Data Collection Websites Megalist (Updated for 2022)
- Appen Connect (Signup Link): Appen has a lot of data collection opportunities ranging from photos, videos, and audio recordings. Most of them require you to use their mobile app for submissions.
- Lionbridge/TELUS International (Signup Link): Lionbridge has a separate email address ([email protected]) that will mail you from time to time regarding data collection opportunities. The payments are much better than Appen. There are a lot of ways to sign up with Lionbridge. Read the Guide for all the signup links.
- Neevo (Signup Link): Neevo has a lot of data collection projects from time to time, and they have a very robust interface that helps you to easily upload. But they are notorious for rejecting a lot of submissions.
- Clickworker (Signup Link): Clickworker has several data collection opportunities from time to time, though they are very infrequent. You are paid 30 days after the approval of your submission.
- Oneforma (Signup Link): Oneforma has a lot of data collection projects for several countries. Their pay rates used to be good but have been reduced nowadays. They usually approve your submissions a lot more than other platforms.
- Dataforce (Signup Link): A data collection platform by Transperfect, usually has image and video tasks only. Check out their active collections by clicking here.
- WOW Crowd (Signup Link): WOW AI is the newest entrant in this space. It only had one job when I looked into it, but they might increase in the future. Worth keeping an eye on.
Update 2022: Robson and Label Docs have been removed from this list. Robson has had no new (global) tasks since months, and Labeldocs has shut down.
Let me know if there are any new legit data collection platforms you know of.