A Raspberry Pi is All You Want
![Towards Data Science](https://miro.medium.com/v2/resize:fill:48:48/1*CJe3891yB1A1mzMdqemkdg.jpeg)
Usually, we have to gather some knowledge inside a sure time frame. It may be knowledge from the IoT sensor, statistical knowledge from social networks, or one thing else. For instance, the YouTube Information API permits us to get the variety of views and subscribers for any channel on the present second, however the analytics and historic knowledge can be found solely to the channel proprietor. Thus, if we need to get weekly or month-to-month summaries about these channels, we have to gather this knowledge ourselves. Within the case of the IoT sensor, there could also be no API in any respect, and we additionally want to gather and save knowledge on our personal. On this article, I’ll present configure Apache Airflow on a Raspberry Pi, which permits working duties for an extended time frame with out involving any cloud supplier.
Clearly, in case you’re working for a big firm, you’ll most likely not want a Raspberry Pi. In that case, in case you want an additional cloud occasion, simply create a Jira ticket in your MLOps division 😉 However for a pet mission or a low-budget startup, it may be an attention-grabbing answer.
Let’s see the way it works.
Raspberry Pi
What is definitely a Raspberry Pi? For these readers who’ve by no means been all for {hardware} for the final 10 years (the primary Raspberry Pi mannequin was launched in 2012), I can briefly clarify that this can be a single-board pc working full-fledged Linux. Normally, a Raspberry Pi has a 1GHz, 2–4-core ARM CPU and 1–8 MB of RAM. It’s small, low cost, and silent; it has no followers and no disk drive (the OS is working from a Micro SD card). A Raspberry Pi wants solely an ordinary USB energy provide; it may be linked through Wi-Fi or Ethernet to a community and run totally different duties inside months and even years.
For my knowledge science pet mission, I needed to gather the YouTube channel statistics inside 2 weeks. For a process that requires solely 30–60 seconds twice per day, a serverless structure generally is a excellent answer, and we are able to use one thing like Google Cloud Operate for that. However each tutorial from Google began with the phrase “allow billing in your mission”. There may be free first credit score and free quotas supplied by Google, however I didn’t need to have one other headache of monitoring how a lot cash I…