How ChatGPT’s Premium Features are Revolutionising Financial Modelling

Financial modelling has long been a mainstay of the financial industry, essential for everything from risk assessment to capital allocation. Generative Artificial Intelligence (AI) technologies, like ChatGPT, are fundamentally changing financial modelling processes.

At the risk of sounding like an advertisement for ChatGPT, the premium features (e.g. access to custom GPTs, image generation, greater access to the latest GPT versions as well as a greater ability to upload files for advanced data analysis and more) available in the paid version are worth exploring as they are extending these capabilities further.

The Wolfram Custom GPT

The Wolfram Custom GPT allows ChatGPT to query Wolfram Alpha’s computational engine to provide a broader range of numerical and computational solutions. This enhances ChatGPT’s capabilities in several ways, including but not limited to:

Mathematical Calculations: Complex calculations, solving equations and even calculus problems can be handled more effectively.
Statistical Analysis: Facilitates advanced statistical computations, including regression analysis, hypothesis testing and data visualisation.
Unit Conversions: Whether it’s distance, weight, or more complex units, the plug-in can provide precise conversions.
Scientific Data: Users can pull in data on various scientific topics, from elemental properties to celestial bodies.
Financial Calculations: Provides financial calculations for things like compound interest, mortgage calculations, etc.
Data Visualisation: The plug-in can create various types of graphs and plots to better illustrate data.
Historical Information: Users can access information on historical events, figures or data.
Weather and Geography: Retrieve detailed current and historical weather data, as well as geographical data like distances between places.
Text and Data Analysis: Natural language text can be analysed for things like word frequency, and more detailed data analysis can be carried out on numerical data sets.

The Wolfram plug-in significantly augments ChatGPT’s capabilities, allowing it to offer a more detailed and nuanced level of response in specific domains.

Advanced Data Analysis (ADA) Feature

While the free version provides limited access to advanced data analysis and file upload functionality, the paid version allows you to do far more with both of these.

It allows you to upload files (Excel, text, CSV, zip, images) then interact with this data using natural language. Behind the scenes, it writes the Python code required to translate your query into code, you can then use the results of this to generate downloadable PDF, CSV or image files.

If it discovers errors or problems in this process, it is even able to correct itself, try different approaches and generate a solution.

The Advanced Data Analysis feature enables ChatGPT to perform tasks such as:

Generate simple descriptive statistics (e.g. mean, median, mode).
Provide basic visualisations like pie charts or bar graphs.
Offer elementary interpretations of the data you provide.
Identify trends or patterns in datasets.

Apply These Components to Financial Modelling

Here’s how these components can be practically applied to financial modelling.

1. Model Creation

One of the biggest benefits of using the ADA and the Wolfram GPT in ChatGPT is that they enable someone with little or no financial modelling skills to create simple models, input their assumptions, produce forecasts, run scenarios and analyse the results. All this can be done just using natural language; no complex Excel formulas or other code is required.

This doesn’t replace the need for financial modellers however as the models produced are relatively simple. Additionally, the model produced can be an unauditable “black box” unless ChatGPT is also asked to show its workings in deriving the results.

The benefit is that it works well for simple situations and it’s quick and it’s very easy for non-technical users, thereby making modelling more accessible.

2. Data Analysis and Preprocessing

Data Cleaning

ChatGPT’s ADA can generate Python code to clean data. This works well for once-off data cleansing operations.

If you need a repeatable process however, you could simply take the Python code and use this within Power Query inside Power BI or Excel. This provides a very visual way to immediately see the results.

Alternatively, if your data is stored within an Excel file, you could take the Python code and use this directly within Excel’s new Python code interpreter to cleanse your data.

Yet another alternative is to use ChatGPT to write the PowerQuery code required to give you the data you need in a specified format.

We’ve also provided some custom GPTs that are specifically tailored to help with writing, cleaning and documenting PowerQuery code, DAX formulas and Excel formulas. Details.

Identify Model Variables

ADA can identify and even create new variables that increase a model’s predictive power. This could be further augmented by Wolfram’s computational algorithms if required.

3. Forecasting

Forecast Method Recommendation

ChatGPT’s ADA can analyse data and recommend a suitable forecasting method. The modeller can then use their judgement to assess whether this is appropriate.

Time-Series Analysis

Advanced forecasting technologies like LSTMs and ARIMA become even more robust when coupled with Wolfram’s computational capabilities. The Python libraries in ADA can also be used to apply forecasting techniques to your data and produce a downloadable dataset.

The dataset produced could then be uploaded into specialist budgeting and forecasting software such as Solver.

Scenario Analysis

Uploading your model’s scenario outputs to ADA allows you to perform richer, more detailed scenario analyses.

4. Simulation

Monte Carlo Simulations

The Wolfram plug-in can be used to perform basic Monte Carlo simulations. If you need more powerful simulations however you would be better to use R or Python libraries then upload the results of these to the ADA for analysis.

If you’re not sure how to do these kinds of things, ChatGPT can provide guidance here also.

5. Decision Support

Insights and Recommendations

For those who aren’t tech-savvy, the ADA makes it easier to analyse large amounts of data to identify trends and relationships. It can also suggest reasons why these trends may have occurred.

Because ChatGPT has such a large base to draw from, the suggestions and recommendations it provides often include many things a single individual wouldn’t come up with on their own.

More accessible, efficient and insightful approaches to data analysis

Using ChatGPT’s premium features, financial modelling becomes easier, more precise and the modeller is able to draw greater insights from a wider database.

In summary, the premium features of ChatGPT are enabling more accessible, efficient and insightful approaches to data analysis, forecasting and decision-making. As these features continue to evolve, they hold the potential to completely reshape how we navigate the complex world of financial modelling.