Vishnu Bharadwaj
Share
  •  
  •  
  •  
  •  

Unstructured data offers a reservoir of insights, which when tapped into, can help investment firms in driving competitive differentiation. In today’s digital era, not only does information impacting stock prices flow in real time, but the sources of data have also multiplied, with several new alternative sources coming to the fore. These include social media posts, geo-location data, and sensor data from IoT-enabled smart devices.

Unstructured Data Analytics _Featured

How this wealth of information can help you boost your business is hardly debatable. Investment firms are sure to gain immensely by running advanced analytics on data from non-conventional sources. Let’s take a look at why updating your investment strategy based on insights from unstructured data is no longer a nice to have; it’s a business imperative.

Alternative sources of data are growing in relevance

The power of information flowing from alternative data sources lies in the fact that they are available in real-time. In the case of stocks, capturing real-time information to fine tune investment strategies can be a game changer. Two key alternative data sources include:

Social media and web traffic

Facebook, Twitter, and other social media platforms are great sources for gathering customer feedback. Performing sentiment analysis on data gathered from these channels can help you predict product demand and performance ahead of product launch. Combined with website traffic volumes, this analysis can prove especially relevant in assessing the impact on the company’s stock price when a new product or service is launched or a major upgrade is planned.

Geo-location data

Analysis of satellite image data can reveal what activities are going on and at what pace. For instance, Chicago-based firm Remote Sensing Metrics provides satellite image data on occupancy rates of parking lots as well as factory activity data from manufacturing units, to its customers – mostly hedge funds. This helps them predict the quarterly performance of these listed companies.

Going from data to decisions

Two of the most commonly used approaches in deriving value from data are:

Rules-based approach

This approach involves making buying and selling decisions based on signals received from the data sources. This can be easily implemented by short-term traders. For instance, T-3, an ad agency has developed a Trump and Dump bot – a robot that algorithmically analyzes the sentiments expressed in U.S President Donald Trump’s tweets. It identifies companies that the president has a negative opinion about and initiates a short-sale for their stocks.

Systematic combination of data sources

This approach involves adding data from new sources alongside the traditional ones for fundamental analysis, and assigning appropriate weights in the overall scoring mechanism used for investment decision making. For this approach, alternative data can be used after making appropriate refinements.  For instance, current traffic data can be combined with property development and parking lot data to forecast real estate demand and price trends in a particular city.  These estimates can be used to make medium term location specific sales forecasts for retailers.

This is more relevant to long-term investors.

Unstructured data can mean big money

Even though alternative data usage is still in its infancy when it comes to mainstream adoption, it is likely to play a big role in investment decision making in the near future. While unstructured data from alternative sources offers critical insights, it is crucial to remove noise and pick only the relevant information. You can achieve this by quantifying relevance of the new sources as opposed to the traditional ones, and establishing a good feedback loop to update the models. Removing correlations among the factors used and understanding the lag effects of some factors on others is also critical. For instance, low parking space occupancy rates (the new source) will also reflect in the traditional source -sales figures (the traditional source) of the retailer’s next quarter’s results. Thus, retailers must not double count. Finally, while comparing stocks and selecting the right ones, it must be ensured that data quality is similar for all companies within a sector to ensure apple-to-apple comparisons

When done right, unstructured data throws up a multitude of opportunities to help you outperform the competition and create value for customers. Well, that’s what we feel. What are your thoughts? Do you think unstructured data analytics will go a long way in helping you provide better investment advisory to your customers? If yes, is your organization ready to take the plunge?


Share
  •  
  •  
  •  
  •