It has been almost seven months since Covid-19 began sweeping through the globe, affecting millions of lives and bringing a level of global disruption unseen in most of our lifetimes. That being said, while we are still some way from a realistic end to this crisis, we are starting to see some of the positive technological developments that have risen in the process of tackling this challenge.
The use of numbers, charts and analytics to inform public, business and government decision-making throughout the pandemic is indicative of the pivotal role data has played in mitigating the threats presented by Covid-19.
Data literacy, collection and analytics has advanced in many areas throughout this time, with increased quantitative information in the mainstream media informing and directing the public amid difficult times. Data analytics has also progressed in business contexts, helping employers to better understand the needs of their organization, pinpoint specific struggles faced by their workforce and plan ahead for an uncertain future.
In light of this, now is the perfect time for reflection to better understand the role data analytics has played throughout the pandemic, the importance of adopting a data-driven approach at this critical time of recovery and the dangers of using unsuitable data sources.
1. Know the context of your data
Making sense of data and using this to mitigate the challenges brought about by Covid-19 starts with understanding the context of the data that has been collected. Raw data that is collected forms the foundation of any subsequent analysis and the quality of the data anchors impact the trustworthiness of the conclusions drawn. To put it simply, ‘rubbish in, rubbish out’.
A significant example of an unreliable data source is the recent paper retraction scandal with The Lancet. The paper claimed that using Hydroxychloroquine on Covid patients increased heartbeat irregularities and death rates, which resulted in serval major Hydroxychloroquine trials being halted. The Lancet then retracted the paper a few days later, with its authors no longer able to “vouch for the veracity of the primary data sources”. This had a devastating impact on the research and treatment landscape for Covid-19 and forces us to consider how many more flawed analyses are out there.
To avoid making the same mistake again, extra time must be taken to better understand the source and collection process of the data being analyzed, whether it is system collected, reported or survey based. Most importantly, once data collection has been grasped – bearing in mind there will always be some level of ambiguity, uncertainty and noise – we must then consider how to clean and process the data before analysis, and whether it places any implication to the interpretation and conclusions drawn.
Ultimately, the pandemic has shone a light on the need for accurate and reliable data and the importance of data-driven decisions that are made based on analyses that are scientifically rigorous and robust.
2. Speed to insight – the availability and timeliness of data
It goes without saying that the availability and timeliness of data and insights are critical when preparing for the unknown. As many have stated, the world is learning at a speed that we have never seen before and the pace of developments is relentless. As a result, in recent months the government and organizations have been compelled to make data-driven decisions within hours or even minutes, not days and months.
To minimize the risk of missed opportunities by failing to capture them in time, swift action must be taken before the value of data diminishes. At present, economists are taking high frequency or real-time data such as job postings and weekly unemployment claims as guidance, due to indicators like GDP tending to have lags. Organizations, healthcare services and government bodies must make informed decisions based on insight from data they have now if they are to get ahead of future challenges.
The bottom line here is that collaboration and speed to insight have been hugely beneficial in overcoming the challenges brought about by Covid-19 and will continue to be in the months ahead.
3. Not all data science is about machine learning and AI
In the current time when the volume of data is exploding and computing power is becoming cheaper every day, data science and machine learning play a leading role in the use of data, elevating organizations from basic level descriptive and diagnostic analytics, to complex predictive or even prescriptive analytics. This enables businesses to learn quickly, plan ahead amid difficult times and, most importantly, take immediate action. Whether this is for supply chain planning, marketing budget allocation, staff resourcing, talent management or customer retention, organizations across all industries are eager to learn how machine learning can improve efficiency across the board.
Throughout the past six months, naturally many wanted to utilize the power of machine learning to help them learn as quickly as possible to ensure preparations could be made for the near and mid-term future. However, during this unprecedented time, we may need to take a step back and re-evaluate our approach. The idea of machine learning is to let the machine learn from existing or historical data. If we unpack this, it means this technique is useful when you have a sufficient amount of data. When events like Covid-19 strike, there might not be enough historical data or past patterns to learn from. This was especially true in the early days of the crisis when analysts were forced to start from scratch.
To combat this issue going forward, there are two viable solutions. The first is to stick with machine learning and use other historical events as a proxy. The second is to consider alternative techniques such as scenario modelling, which is a process of examining and evaluating possible events that could happen in the future. Either way, finding new ways to conquer data and model the unknown – coupled with the ongoing commitment from organizations to learn and grow in an agile manner – will be pivotal to their success in the coming months.
4. Data visualization ≠ storytelling.
Charts and graphs can be seen anywhere and everywhere these days. People are re-sharing the numbers they read or see in the news, in both informal and formal ways. Data analysis and visualization skills have been largely democratized, with desktop software such as Tableau and online visualization tools such as Google Data Studio and Chartbuilder widely accessible. To put it simply, data literacy has become one of the most sought-after skills for employers.
Then again, data literacy is not just about creating beautiful charts or coding complex functions to wrangle data. Analyzing data – apart from cleaning, transformation and visualization – means interpreting the results and gaining insight from lines, bars, maps, numbers, shades or whatever format might be presented. Visualization does not equate to interpretation and storytelling, which often requires subject matter expertise or years of industry experience. Ultimately, analysts must delve deep into what data can offer and how this might help improve business processes, efficiencies or any desirable outcome by informing strategic and operational decisions.
5. It’s all about actions
And finally – perhaps the most important factor to remember – driving insight from data is all about actions. The majority of an analytics team’s time and effort are typically on data collection, cleaning and visualization, but the biggest impact is made by interpretation and action. When a piece of analysis is finished, it is important that the analytics teams are thinking about what is next and how the insights feed into the business as it is easy to forget the primary goal of data analytics when buried deep in analysis.
If organizations are willing to get on the front foot in terms of data literacy, collection and data science, there is every chance they will continue on this road of recovery and survive or even thrive in the challenging times ahead. Innovation, speed and collaboration will be key, and organizations in every sector should now be focusing their efforts on how to utilize data to navigate the uncertain months to come.
Binqian Gao, Data Science Lead, TrueCue