Augvest hosted events in Singapore and Hong Kong in early Sep 2016 to connect with funds and understand the current state of data driven investing in the region - and how it is likely to develop.
In total, we spoke with some 35 funds. These included bottom-up investors, quant funds, and a few hybrids.
While the use of "alternative data" at some funds in mainland China is easily on par with what we have seen among market leaders in the US, many firms based in Hong Kong and Singapore are only beginning to become familiar with the term.
Most conversations began with us defining "alternative data" as follows:
What it's not
- Traditional financial datasets like returns / trade volumes and financial statement data.
- Outputs from Natural Language Processing (NLP) - heatmaps, sentiment scores, event alerts.
- Data generated inside your fund - your past positions, your P&L, etc.
What it is
- Data from service providers like payment processors, various consultants, web analytics firms, etc.
- Selections from large-scale data ingestion - e.g. from large-scale images from satellites, from IoT, etc.
- Systematically archived public data - e.g. from company websites or from government records.
Other important elements to touch on are:
Generally speaking, there are two models. Targeted sourcing involves identifying important KPIs, figuring out who might have relevant data, and then initiating discussions with those firms. Opportunistic sourcing involves running an ongoing search for data on any and all of the companies in the fund's investable universe.
Building out ETL (Extract / Transform / Load) pipelines is an underappreciated aspect of alternative data - it is often a far greater challenge than the statistical modeling that follows.
Once data has been ingested and is prepared for analysis, your data analyst has to extract key features, use them to build a model, and then determine how best to present the output. All three of these steps require intuition around the fundamental analysis within which the data will be used.
Use cases funds in Asia found particularly interesting include:
- Several businesses in the US sell revenue estimates sourced from credit card data collected by a major payment processing platform. Market leaders pay a much higher price to gain access to the raw source data and they have found that they can do a much better job cleaning and processing it.
- Ad exchanges collect data on the devices through which ads are to be displayed. All Teslas come equipped with distinct tablets; counting the number of such tablets allows you to track numbers of Teslas delivered.
- One of the major new entrants into the alternative data space is Singapore's sovereign wealth fund. Roughly 10% of all maritime trade passes through Singapore's ports; the government is able to access all of that port data and use it to track both individual companies as well as macro trends.
If you are looking for data or are considering hiring a data scientist - we may be able to help. Please email us.