top of page
  • Writer's pictureGunjan Syal

Suspect Data Analytics

Updated: Nov 22, 2022

Data Analytics Charts
Media From Wix

I have been binging on Mona Chalabi's TED Talks lately. Mona is an accomplished data journalist and has an inspirational way of bringing the data closer to the topics of real social impact. I especially enjoy it when she is joined by a puppet representing her mother's accomplished spirit. Mona shares the term 'suspect statistics' in one of her TED Talks. The topic is spotting bad statistics. This is not necessarily malicious statistics put together to mislead people; it could be statistics that are easy to misinterpret or lack clarity. The end result is distorted representation of facts.

This had me reflecting on my own 16+ years of experience enabling business and technology projects. Many of them involved the use of data to make key organizational decisions. These decisions often involved multiple data sets and required many layers of analysis by various business/ external groups. Then, the results could be aggregated and presented to the decision-making groups such as executives and the boards. Inspired by Mona's 'suspect statistics', a new term 'suspect data analytics' popped up in my mind.

Suspect data analytics is not necessarily data prepared with the explicit purpose of misleading or misrepresenting. Results from data analytics become suspect over time, simply because multiple layers or volumes of poorly defined data are aggregated on top of one another until the context is lost or distorted. This is much like a game of telephone the children play, when passing messages to each other by whispering.

Ultimately, the final result and deliverables may seem to make sense in accordance with the general algorithms or processes, however fail to drive forward decisions of any real business value. This leads to failed investments in form of negative return on investments (ROI) from data projects, that are not recognized by business leaders until it is too late.

The key to avoiding suspect data analytics involves four principles:
  1. preparing and articulating an enterprise-wide data strategy,

  2. asking the right questions during the data collection,

  3. documenting the organizational context with the data analytics, and

  4. sharing it in timely manner with the right audience.

This is our area of specialty at GoEmerald. We look at data from an enterprise-wide perspective to enable the long-term ROI in the end; both for the organization and the society with the responsible innovation considerations. If you are interested in learning more about our services and work, contact us for an appointment, or sign up below for future event notifications.

Upcoming Events

If you enjoyed this post, join the global transform this community for future editions.


© 2022 Emerald Technology Group Inc. All rights reserved.

This work and the entirety of its contents, including all images, text, frameworks, articles and operating names, hereinafter referenced as "Proprietary Property", are protected under copyright, trademark, patent and other laws in Canada and/or other countries. All rights, title, and interest in the Proprietary Property are solely owned by Emerald Technology Group Inc. No License (direct or implied) is granted to anyone by this transmission of the Proprietary Property. Nor shall such a transmission constitute any representation, warranty, assurance, guaranty or inducement by the owner of the Proprietary Property to anyone with respect to infringement of intellectual property rights of others. It is not permitted to make any copies or reproduction of the Proprietary Property in any form including digital, print or otherwise, translate, edit, alter, modify, or create any derivative or similar works of the Proprietary Property without express written permission from the owner of the Proprietary Property. For permissions, please contact:


Commenting has been turned off.
bottom of page