Ana-Lyza offers low latency, easy access to financial data

Ana-Lyza reduce costs and accelerate time-to-market in the ETL of financial data, unlike time-consuming traditional methods that are prone to human error.

Comment 5

Inspiration

Ana-Lyza came from inspiration to contribute women-led fintech solutions to the industry. Analyza's ultimate goal is to empower women and marginalised community who has no investing background as well as promoting ESG and Gender Lens Scoring in investments. According to Deloitte, the global fintech founder community is still dominated by men, with women making up just 7% of the total pool. The fintech products out there are made mostly by men for men. Forbes' recent study shows men (23%) are investing in the stock market more since Covid-19 compared to only 10% of women.

From my experience as a financial journalist and editor, interviews with early adopters of this tech, and external research on tech for the financial sector, there is a pressing need for a more accessible tech that helps financial analysts, publications, fund managers, and investors to cut time and costs performing ETL tasks from documents to financial data.

Gartner forecasts elimination of mundane tasks inevitable

Gartner in its latest research forecasted that by 2023, there will be less reliance on analysts for repetitive and mundane tasks such as ETL of data from financial statements. Gartner said this could free up to 20% of their time for collaboration, training and high-value analytics.

By 2023, 95% of Fortune 500 companies will converge data and analytics, said Gartner. It expects broader initiatives encompassing data and analytics governance initiatives.

Persona interview shows pain points

Additionally, I had the opportunity to interview a few personas during the development of this project. The persons, who are early adopters of the tech, shared their pain points as follows:

1) Two out of three extract historical data manually: This manual method is very time consuming not to mention there are high risks of human error 2) Two out of three hire third-party services: Not only it is costly but frustrating and time-consuming as half of the time analysts/fund managers have to correct human errors and do re-entry of data 3) Investors in general lack access to standardized ESG and Gender Lens Investment data : Investors are not able to make informed decisions unless they read multiple sustainability reports or/and media articles to check if the company's operations truly align with their ESG and Gender Lens focus

Opportunity to tap into small to medium financial services

There is currently a vacuum for medium to small research houses, publications, and fund management units that are less inclined to spend on expensive third-party services that use traditional methods or subscribe to costly services such as the Bloomberg Terminal.

What it does

Ana-Lyza empowers users to extract and analyze data at a fraction of the cost versus other in-house, traditional, and manual ETL methods. The features of the app are:

1) ETL of financial data (general information, revenue, net profit&loss, cash flow) from 10 K-Form (PDF Format). The segments analyzed are pages from the cover page up to page 30 and financial statements, using Azure Form Recognizer API

2) Text summarization or Key Phrase Extraction from the respective company's 10K Business and/or Risk Factors segments, using Azure Cognitive Services (Language) API

3) Automated financial model to calculate and display revenue/income growth as well as complex financial ratios to help investors make informed decisions. Financial Ratios include Earnings per share (EPS), Debt-to-equity ratio, and Current Ratio.

4) ESG Scoring based on the company's sustainability report sourced from https://www.responsibilityreports.com. This feature uses zero-shot classification(Hugging Face, Microsoft DeBert Model) at Azure ML notebook. At the time of submission Hugging face partnership with Azure was still very new and i didnt have the time to redo the entire solution.

How I built it

Ana-Lyza was built with Azure Form Recognizer, Azure Cognitive Language (Services), Azure Machine Learning Studio, Azure Blob Storage, Azure Active Directory, and PowerApps.

Challenges I ran into

Beginners errors and limited resources

It took me a bit of time (and credits) when deploy Azure services. For example, using the right region, and training data by tagging documents. I also had to migrate halfway to using Form Recognizer 3.0 vs From Recognizer 2.0.

I was also challenged with the limited resources (provided for F0 tier users), for example when training data using Form Recognizer, the collection and training of data was limited to only 500pages.

Integrating Azure with PowerApps

I met with challenges when calling the API for example, getting the right response to the request body when establishing a custom connector.

Accomplishments that I'm proud of

Despite the limited resources and beginner errors, I was able to develop the desired application to perform . What's more, I was able to leverage Power Apps and Azure features to add additional features to make the app more desirable and scalable.

What I learned

During the month-long hackathon, I learned to solve problems better by raising tickets at Azure, Stack overflow, and sourcing the right Documentation. I have enhanced my knowledge and skills in building Machine Learning (ML) models and how ML can be applied to automate and enhance current traditional business processes. I also learned how to better deploy Microsoft Azure as well as Microsoft's other suite of products such as Power Apps.

What's next for Ana-Lyza

The wish list for Ana-Lyza's future development include a robo investment advisor promoting investment in ESG and Gender Lens performing companies, while taking into fundamentals based on the company's earnings/revenue growth, risks, etc. It is also possible to include predictions of the company's financial standings as well as stock prices. Ana-Lyza should also be able to accept other formats of input including XBRL/spreadsheet as the Form Recognizer capabilities are being enhanced. That said, Ana-Lyza could also be customized based on the different IFRS accounting/ reporting standards of different countries and regions.

Go to market Strategy

We target users (analysts, fundmanagers,investors,publications) who are less inclined to subscribe costly 3rd party services such as Bloomberg etc. We also find opportunities for tertiary students who can use the service for assignments and thesis.

After a swot analysis we propose: 1)subscription for Basic and Premium services (Using Azure pricing calculator) of USD69 permonth and USD 99 per month respectively. for prebuilt models and other bespoke services 2) customised models and services are also available for medium to large businesses

Built With

+ 4 more
Share this project:

Updates