Fast Forecasting Tools on Top of Google Sheets

Google Sheets is often the bread and butter of internal operations. Teams keep track of monthly revenue and how much every customer pays the company. Teams keep track of expected and projected utilization of human resources. Ride sharing companies may keep track of how many contractors they need to employ per project per a quarter. ML companies may keep track of how many annotators they need to help label their datasets.

However, forecasting into the future is often done manually–and in an ad-hoc fashion. Although you may want to forecast various time series, it is challenging to explore all factors contributing to the predictive power of your system without hiring an expensive data scientist or engineer.

If you're working with complicated datasets that are scattered across siloed systems, there may be additional engineering challenges to bring this system together in one place.

One possible solution is a forecasting dashboard (like the one below), that reads data from your Google Sheet and is able to give you predicted outputs, as well as uncertainty estimations and error scores for your system. Having to build such a system from scratch may require a lot of boilerplate code, components and time.

Today, we're excited to announce DataSiv's automated machine learning framework that enables you to quickly spin up forecasting tools on top of any data source. Our framework has predictive power with limited labeled data and is able to take into account seasonality, as well as correlation across time series for a unified model.

Best of all, it is very fast to set up (spinning up this dashboard took less than 30 minutes) and only requires one-time-setup to reuse every month, week or hour.

If you're interested in spinning up internal forecasting tools on top of any data source such as Google Sheets, AirTable or a SQL database check out our tutorial here and sign up for an account.