Vibe Coding a Non-public AI Monetary Analyst with Python and Native LLMs

Picture by Writer

# Introduction

 
Final month, I discovered myself watching my financial institution assertion, making an attempt to determine the place my cash was really going. Spreadsheets felt cumbersome. Present apps are like black containers, and the worst half is that they demand I add my delicate monetary knowledge to a cloud server. I needed one thing completely different. I needed an AI knowledge analyst that would analyze my spending, spot uncommon transactions, and provides me clear insights — all whereas preserving my knowledge 100% native. So, I constructed one.

What began as a weekend challenge was a deep dive into real-world knowledge preprocessing, sensible machine studying, and the facility of native massive language fashions (LLMs). On this article, I’ll stroll you thru how I created an AI-powered monetary evaluation app utilizing Python with “Vibe Coding.” Alongside the way in which, you’ll be taught many sensible ideas that apply to any knowledge science challenge, whether or not you’re analyzing gross sales logs, sensor knowledge, or buyer suggestions.

By the top, you’ll perceive:

• Easy methods to construct a sturdy knowledge preprocessing pipeline that handles messy, real-world CSV recordsdata
• How to decide on and implement machine studying fashions when you’ve got restricted coaching knowledge
• Easy methods to design interactive visualizations that really reply person questions
• Easy methods to combine a neighborhood LLM for producing natural-language insights with out sacrificing privateness

The entire supply code is out there on GitHub. Be at liberty to fork it, prolong it, or use it as a place to begin to your personal AI knowledge analyst.

Fig. 1: App dashboard exhibiting spending breakdown and AI insights | Picture by Writer

# The Downside: Why I Constructed This

 
Most private finance apps share a elementary flaw: your knowledge leaves your management. You add financial institution statements to providers that retailer, course of, and doubtlessly monetize your info. I needed a instrument that:

1. Let me add and analyze knowledge immediately
2. Processed every little thing regionally — no cloud, no knowledge leaks
3. Offered AI-powered insights, not simply static charts

This challenge turned my automobile for studying a number of ideas that each knowledge scientist ought to know, like dealing with inconsistent knowledge codecs, choosing algorithms that work with small datasets, and constructing privacy-preserving AI options.

# Undertaking Structure

 
Earlier than diving into code, here’s a challenge construction exhibiting how the items match collectively:

challenge/   
  ├── app.py              # Principal Streamlit app
  ├── config.py           # Settings (classes, Ollama config)
  ├── preprocessing.py    # Auto-detect CSV codecs, normalize knowledge
  ├── ml_models.py        # Transaction classifier + Isolation Forest anomaly detector
  ├── visualizations.py   # Plotly charts (pie, bar, timeline, heatmap)
  ├── llm_integration.py  # Ollama streaming integration
  ├── necessities.txt    # Dependencies
  ├── README.md           # Documentation with "deep dive" classes
  └── sample_data/
    ├── sample_bank_statement.csv
    └── sample_bank_format_2.csv

We are going to take a look at constructing every layer step-by-step.

# Step 1: Constructing a Sturdy Knowledge Preprocessing Pipeline

 
The primary lesson I discovered was that real-world knowledge is messy. Completely different banks export CSVs in fully completely different codecs. Chase Financial institution makes use of “Transaction Date” and “Quantity.” Financial institution of America makes use of “Date,” “Payee,” and separate “Debit”https://www.kdnuggets.com/”Credit score” columns. Moniepoint and OPay every have their very own kinds.

A preprocessing pipeline should deal with these variations routinely.

// Auto-Detecting Column Mappings

I constructed a pattern-matching system that identifies columns no matter naming conventions. Utilizing common expressions, we will map unclear column names to plain fields.

import re

COLUMN_PATTERNS = {
    "date": [r"date", r"trans.*date", r"posting.*date"],
    "description": [r"description", r"memo", r"payee", r"merchant"],
    "quantity": [r"^amount$", r"transaction.*amount"],
    "debit": [r"debit", r"withdrawal", r"expense"],
    "credit score": [r"credit", r"deposit", r"income"],
}

def detect_column_mapping(df):
    mapping = {}
    for subject, patterns in COLUMN_PATTERNS.objects():
        for col in df.columns:
            for sample in patterns:
                if re.search(sample, col.decrease()):
                    mapping[field] = col
                    break
    return mapping

The important thing perception: design for variations, not particular codecs. This strategy works for any CSV that makes use of widespread monetary phrases.

// Normalizing to a Customary Schema

As soon as columns are detected, we normalize every little thing right into a constant construction. For instance, banks that break up debits and credit should be mixed right into a single quantity column (adverse for bills, optimistic for revenue):

if "debit" in mapping and "credit score" in mapping:
    debit = df[mapping["debit"]].apply(parse_amount).abs() * -1
    credit score = df[mapping["credit"]].apply(parse_amount).abs()
    normalized["amount"] = credit score + debit

Key takeaway: Normalize your knowledge as quickly as doable. It simplifies each following operation, like function engineering, machine studying modeling, and visualization.

Fig 2: The preprocessing report exhibits what the pipeline detected, giving customers transparency | Picture by Writer

# Step 2: Selecting Machine Studying Fashions for Restricted Knowledge

 
The second main problem is restricted coaching knowledge. Customers add their very own statements, and there’s no large labeled dataset to coach a deep studying mannequin. We want algorithms that work properly with small samples and could be augmented with easy guidelines.

// Transaction Classification: A Hybrid Method

As an alternative of pure machine studying, I constructed a hybrid system:

1. Rule-based matching for assured instances (e.g., key phrases like “WALMART” → groceries)
2. Sample-based fallback for ambiguous transactions

SPENDING_CATEGORIES = {
    "groceries": ["walmart", "costco", "whole foods", "kroger"],
    "eating": ["restaurant", "starbucks", "mcdonald", "doordash"],
    "transportation": ["uber", "lyft", "shell", "chevron", "gas"],
    # ... extra classes
}

def classify_transaction(description, quantity):
    for class, key phrases in SPENDING_CATEGORIES.objects():
        if any(kw in description.decrease() for kw in key phrases):
            return class
    return "revenue" if quantity > 0 else "different"

This strategy works instantly with none coaching knowledge, and it’s straightforward for customers to grasp and customise.

// Anomaly Detection: Why Isolation Forest?

For detecting uncommon spending, I wanted an algorithm that would:

1. Work with small datasets (not like deep studying)
2. Make no assumptions about knowledge distribution (not like statistical strategies like Z-score alone)
3. Present quick predictions for an interactive UI

Isolation Forest from scikit-learn ticked all of the containers. It isolates anomalies by randomly partitioning the information. Anomalies are few and completely different, so that they require fewer splits to isolate.

from sklearn.ensemble import IsolationForest

detector = IsolationForest(
    contamination=0.05,  # Anticipate ~5% anomalies
    random_state=42
)
detector.match(options)
predictions = detector.predict(options)  # -1 = anomaly

I additionally mixed this with easy Z-score checks to catch apparent outliers. A Z-score describes the place of a uncooked rating when it comes to its distance from the imply, measured in normal deviations:
[
z = frac{x – mu}{sigma}
]
The mixed strategy catches extra anomalies than both methodology alone.

Key takeaway: Generally easy, well-chosen algorithms outperform complicated ones, particularly when you’ve got restricted knowledge.

Fig 3: The anomaly detector flags uncommon transactions, which stand out within the timeline | Picture by Writer

# Step 3: Designing Visualizations That Reply Questions

 
Visualizations ought to reply questions, not simply present knowledge. I used Plotly for interactive charts as a result of it permits customers to discover the information themselves. Listed below are the design ideas I adopted:

1. Constant coloration coding: Crimson for bills, inexperienced for revenue
2. Context by way of comparability: Present revenue vs. bills facet by facet
3. Progressive disclosure: Present a abstract first, then let customers drill down

For instance, the spending breakdown makes use of a donut chart with a gap within the center for a cleaner look:

import plotly.categorical as px

fig = px.pie(
    category_totals,
    values="Quantity",
    names="Class",
    gap=0.4,
    color_discrete_map=CATEGORY_COLORS
)

Streamlit makes it straightforward so as to add these charts with st.plotly_chart() and construct a responsive dashboard.

Fig 4: A number of chart sorts give customers completely different views on the identical knowledge | Picture by Writer

# Step 4: Integrating a Native Massive Language Mannequin for Pure Language Insights

 
The ultimate piece was producing human-readable insights. I selected to combine Ollama, a instrument for working LLMs regionally. Why native as an alternative of calling OpenAI or Claude?

1. Privateness: Financial institution knowledge by no means leaves the machine
2. Value: Limitless queries, zero API charges
3. Pace: No community latency (although era nonetheless takes a couple of seconds)

// Streaming for Higher Person Expertise

LLMs can take a number of seconds to generate a response. Streamlit exhibits tokens as they arrive, making the wait really feel shorter. Right here is an easy implementation utilizing requests with streaming:

import requests
import json

def generate(self, immediate):
    response = requests.put up(
        f"{self.base_url}/api/generate",
        json={"mannequin": "llama3.2", "immediate": immediate, "stream": True},
        stream=True
    )
    for line in response.iter_lines():
        if line:
            knowledge = json.hundreds(line)
            yield knowledge.get("response", "")

In Streamlit, you may show this with st.write_stream().

st.write_stream(llm.get_overall_insights(df))

// Immediate Engineering for Monetary Knowledge

The important thing to helpful LLM output is a structured immediate that features precise knowledge. For instance:

immediate = f"""Analyze this monetary abstract:
- Whole Earnings: ${revenue:,.2f}
- Whole Bills: ${bills:,.2f}
- Prime Class: {top_category}
- Largest Anomaly: {anomaly_desc}

Present 2-3 actionable suggestions primarily based on this knowledge."""

This offers the mannequin concrete numbers to work with, resulting in extra related insights.

Fig 5: The add interface is easy; select a CSV and let the AI do the remaining | Picture by Writer

// Operating the Utility

Getting began is simple. You will want Python put in, then run:

pip set up -r necessities.txt

# Non-compulsory, for AI insights
ollama pull llama3.2

streamlit run app.py

Add any financial institution CSV (the app auto-detects the format), and inside seconds, you will note a dashboard with categorized transactions, anomalies, and AI-generated insights.

# Conclusion

 
This challenge taught me that constructing one thing purposeful is just the start. The actual studying occurred once I requested why each bit works:

• Why auto-detect columns? As a result of real-world knowledge doesn’t comply with your schema. Constructing a versatile pipeline saves hours of handbook cleanup.
• Why Isolation Forest? As a result of small datasets want algorithms designed for them. You don’t all the time want deep studying.
• Why native LLMs? As a result of privateness and price matter in manufacturing. Operating fashions regionally is now sensible and highly effective.

These classes apply far past private finance, whether or not you’re analyzing gross sales knowledge, server logs, or scientific measurements. The identical ideas of strong preprocessing, pragmatic modeling, and privacy-aware AI will serve you in any knowledge challenge.

The entire supply code is out there on GitHub. Fork it, prolong it, and make it your individual. In case you construct one thing cool with it, I’d love to listen to about it.

// References

Shittu Olumide is a software program engineer and technical author captivated with leveraging cutting-edge applied sciences to craft compelling narratives, with a eager eye for element and a knack for simplifying complicated ideas. You can too discover Shittu on Twitter.