Data is an essential ingredient for spatial analysis—which is predicated on access to your own data, plus useful third-party data. In the end, spatial analysis is about putting your data in the context of its surroundings, creating a greater understanding of your own geographies.
However, the main problem in data analytics today is the amount of time Data Scientists spend gathering the right data. We frequently hear that 80% of their total analysis time is dedicated to gathering, cleaning and feature preparation, while only the remaining 20% is actually spent on analysis, modeling, and communication of results. In the case of Spatial Data Science this is an even bigger issue, as gathering data requires working with highly varied sources.
Here at CARTO we’re on a mission to increase the efficiency of Data Science teams who need to work with spatial data and create spatial models.
We’re thrilled to announce a private beta release of Data Observatory 2.0, the latest version of our spatial data repository that enables Data Scientists to augment their data and broaden their analysis with the latest and greatest in location data—such as human mobility, weather, and financial transactions data.
This new version of Data Observatory will help Data Scientists to save time through an improved data experience via:
Spending 80% of your time gathering and cleaning data is wasteful, making your entire Data Science workflow longer and more tedious. Such a lengthy process also increases the probability of making mistakes, that can put the validity of an entire model at risk. In many cases we see organizations not using third party data altogether because of the hassle.
We found three key issues hindering the data gathering process:
If an analysis took, for example, one full month to get done, all three of these issues probably required around three weeks of work.
The result of this situation is not only frustration and a slowdown in the creation of analysis, but also many organizations give up licensing third party data because it’s just too expensive. And not because the data is expensive, but because the resource cost of acquiring it and making use of it can be far too high.
Data itself is not expensive—accessing and using it is, and the root cause for this is how data is delivered.
Let’s take, for example, a Data Scientist who is working on a market analysis in Portugal for retail expansion. She will start looking for demographics, purchasing power, geographies, etc. Let’s breakdown the costs (based on our own customers’ experience):
Data is not the most expensive piece by any means, it is merely the overhead! This is reflective of the current drama of the data industry.
If we can reduce the cost of getting access to and using the data, more data will be used and purchased, and therefore the data cost will be reduced, which will results in more data being used, and so on, in a continuous virtuous cycle.
CARTO’s Data Observatory 1.0 already tackled some of these challenges, but we saw an opportunity with Data Observatory 2.0 to think more holistically. Given all the broken processes in the location data industry, it was time to rethink it from the ground up. With the release of Data Observatory 2.0, we’re introducing a new experience that will allow Data Scientists to unlock the power of spatial analysis and new location data streams.
The first type of user we want to support with this new version of the Data Observatory is the Data Scientist, and we all know that most of them live in Python Notebooks these days. Therefore we want to provide them with the possibility to manage data discovery, purchase, and access without having to leave their analysis environment. From the launch date, CARTO’s Data Observatory 2.0 is fully accessible within a Jupyter Notebook or Python script, so there is no need to switch platforms.
The following are the core principles that have guided us throughout the process of rebuilding the Data Observatory:
The new Data Observatory 2.0 is accessible via CARTOframes, our Python package, to selected CARTO Enterprise users in a private beta. We’re still open to more beta testers, so if you’re interested, please get in touch.
If you are a data provider interested in adding your datasets to Data Observatory 2.0, we’d also like to hear from you - please contact us.
Want to understand how it would work with your data?Request a demo!
As announced at Google Cloud Next ‘20, we are very pleased to provide more details about our BigQuery Tiler which is now available as a beta.News
In our continuous efforts to make sure CARTO works well with other technologies, we’re always assessing how we can improve our customers’ workflows so that they can easily ...News
Please fill out the below form and we'll be in touch real soon.