In our latest round up of data-related news and views, we look at the influence of AI on carbon emissions, how spreadsheets have helped shape the modern economy (but pose major business risks) and more.
Achieving Zero Greenhouse Gas Emissions by 2050
2050 might be too late for life as we know it, but this interesting infographic from Information is Beautiful breaks down the changes needed to make that target a reality. See the full graphic here.
How Spreadsheets Helped Make the Modern Economy
This BBC podcast looks back at the invention of the digital spreadsheet and how it quickly transformed accountancy and financial planning and management. Its impact has spread to every aspect of modern business.
Salesforce Acquires Tableau for $15.7 Billion
There are only a few technology systems that evoke passion in their users. Tableau is one of them.
It allows data, from a variety of sources, to be dynamically displayed and used graphically by users. Very powerful. And, Tableau was just acquired by Salesforce. This is part of a trend towards the visualization of data by large tech platforms.
Read more on the acquisition at Geekwire here.
Let the Spreadsheet Do the Mathematics – You Focus on the Story
Many journalists see themselves primarily as storytellers and word artists – numbers can be worrisome, boring or interfere with the flow of an article.
This article from Global Investigative Journalism Network (GIJN) discusses how reporters can let spreadsheets do much of the data crunching work, resulting in more meaningful, insightful journalism.
Deep Learning Has a Terrible Carbon Footprint
There’s a common misconception that computing and web-based technologies are virtually free of environmental impact. After all, if we conduct our meetings via Skype there’s no need for travel, and if we communicate by email there’s no waste paper – right?
It’s often forgotten that data centers can be hugely energy-intensive and the computing power required for artificial intelligence can be massive.
Researchers at the University of Massachusetts, Amherst, assessed the energy needed to train several common large AI models. They found that the process can emit over 626,000 pounds of carbon dioxide equivalent, which is nearly 5X the emissions of the average American car, including the manufacturing process, over its entire lifetime.
The team also found that this equates to 300X the CO2 emissions of a round trip flight from New York to San Francisco. Given the scale of the environmental impact of AI / machine learning, we clearly need to consider how we can marry the goals of reducing emissions with the growing influence of these technologies on our every day lives.
Read an MIT Technology Review article on the subject here and the original research paper here.
Spreadsheet Risk is Real – Yet Businesses are Ignoring It
A study conducted by Forrester Consulting and commissioned by Incisive Software found that “C-level executives are making decisions based on data assumed to be accurate, but that can contain errors” .
Frontline workers recognize the issue, but the study found that “nearly a third of respondents noted that management doesn’t recognize this risk”.
Fewer that 20% of the respondents from 170 mid to large North American businesses felt sufficiently empowered to deal with the problem.
Read more about the study on this Yahoo! Finance article.
Learn about EQUS – our useful Excel plugin for formula visualization, helping spreadsheet users to identify potential errors.