Update README.md

README.md

@@ -7,91 +7,99 @@ sdk: docker
 app_port: 7860
 pinned: false
 short_description: CAPM + CML portfolio advisor w/ embeddings
-license: cc-by-4.0
 models:
   - FinLang/finance-embeddings-investopedia
 datasets:
-  - your-username/LensIQ-Synthetic-Portfolio-1000
+  - Tulitula/LensIQ-Synthetic-Portfolio-1000
 ---
 
 # LensIQ — CAPM + CML Portfolio Advisor
 
-LensIQ is a Gradio app for portfolio analysis using CAPM and the Capital Market Line (CML), with embedding-aided suggestions.
+- **Plot:** x = historical σ (annualized, from monthly covariances), y = CAPM E[r] = rf + β·ERP.  
+- **Efficient mixes:** market/bills with (i) same σ, (ii) same E[r] as your portfolio (weights shown).  
+- **Suggestions:** one Low/Medium/High, from 1,000 long-only mixes built **from your tickers** (+ VOO), ranked with CAPM + embeddings (MMR).  
+- **Export:** every run writes a CSV of all 1,000 candidates in `data/`.
 
-- **Plot**
-  - **x-axis**: historical σ (annualized, from monthly covariance of your tickers)
-  - **y-axis**: CAPM E[r] = rf + β·ERP
-  - CML drawn from risk-free to market (**VOO**).
+> Include **VOO** in your portfolio before computing.
 
-- **Efficient mixes** (market/bills only)
-  - **Same σ** as your portfolio → point on the CML at your σ
-  - **Same E[r]** as your portfolio → σ on the CML for that return
-  - Weights are shown so you can replicate the mixes.
+## Run Locally
 
-- **Suggestions**
-  - Exactly **one** candidate per band (**Low / Medium / High**).
-  - Built from **1,000** random long-only mixes using **your tickers** (+ `VOO` as market proxy).
-  - Ranking uses CAPM metrics + **FinLang/finance-embeddings-investopedia** with MMR for diversity.
+```bash
+docker build -t lensiq .
+docker run -p 7860:7860 -e GRADIO_SERVER_NAME=0.0.0.0 lensiq
+# open http://localhost:7860
+Create & Upload the Dataset
+In the app, add tickers (incl. VOO), set amounts, horizon, lookback → Compute.
 
-- **Export**
-  - Each run saves a CSV with all candidate mixes (downloadable from the UI).
+Download the generated CSV from the UI (saved under data/).
 
-> **Note:** Include **VOO** in your portfolio (market proxy) before computing.
+Create a dataset repo on Hugging Face named Tulitula/LensIQ-Synthetic-Portfolio-1000 and upload:
 
----
+the CSV you downloaded
 
-## Run Locally
+the dataset README from below (Section B).
 
-    docker build -t lensiq .
-    docker run -p 7860:7860 -e GRADIO_SERVER_NAME=0.0.0.0 lensiq
-    # Open http://localhost:7860
+swift
+Kopiér
+Rediger
 
----
+**B) Dataset README (for `Tulitula/LensIQ-Synthetic-Portfolio-1000`)**
+```markdown
+# LensIQ Synthetic Portfolio 1000
 
-## Using the App
+Synthetic portfolio dataset generated by the LensIQ app for CAPM/CML-based analysis and ranking.
 
-1. Add tickers and include **VOO**.  
-2. Enter **$ amounts** for each ticker.  
-3. Set **Horizon** and **Lookback** (years).  
-4. Click **Compute**.  
-5. Review your portfolio, the two efficient mixes (with weights), and the **one** Low/Medium/High suggestion (with % and $).  
-6. Download the generated CSV.
+- **Size:** 1,000 rows (per export)
+- **Modality:** Tabular (text fields for lists)
+- **License:** CC BY 4.0
+- **Intended use:** Research/teaching on portfolio construction, risk/return modeling, ranking/retrieval. *Not financial advice.*
 
----
+## Generation Process
 
-## Method
+Given a user universe **U** (their tickers + `VOO` as market):
 
-- Prices: Yahoo Finance monthly (`yfinance`)  
-- Risk-free: FRED tenor by horizon (1–30y)  
-- Market: `VOO`  
-- ERP: market CAPM E[r] − rf  
-- β: covariance vs. market (excess returns)  
-- σ (hist): annualized from monthly covariance  
-- Suggestions: 1,000 long-only mixes; embedding-aided ranking + MMR
+1. Sample `k ~ U{2…min(8, |U|)}` tickers from **U** (no replacement).  
+2. Draw long-only weights `w ~ Dirichlet(1,…,1)`.  
+3. Compute:
+   - Prices: Yahoo Finance monthly (auto-adjusted), last *N* years (lookback).
+   - Risk-free `rf`: FRED tenor chosen by horizon (e.g., DGS10).
+   - Market `m`: `VOO`.
+   - Market ERP `= E[m] − rf` (from monthly returns, annualized).
+   - Asset betas vs. market (excess returns).
+   - **mu_capm** `= rf + β_p·ERP`, where `β_p = Σ_i w_i β_i`.
+   - **sigma_hist**: annualized portfolio σ from monthly covariance.
+   - **sigma_capm**: `|β_p| · σ_market` (for reference to CML).
+4. Rank candidates with CAPM metrics + finance embeddings (MMR) to surface Low/Medium/High picks in the app.
 
----
-
-## Synthetic Dataset
+## Columns
 
-A CSV is written under `data/`, e.g. `data/investor_profiles_<timestamp>.csv`.
+| column       | type   | description                                            |
+|--------------|--------|--------------------------------------------------------|
+| `tickers`    | string | comma-separated tickers in the candidate mix          |
+| `weights`    | string | comma-separated floats summing to 1 (long-only)       |
+| `beta`       | float  | portfolio beta (weighted by `weights`)                |
+| `mu_capm`    | float  | CAPM expected annual return                           |
+| `sigma_hist` | float  | annualized historical σ from covariance               |
+| `sigma_capm` | float  | CAPM σ on the CML for equivalent return               |
 
-**Columns:** `tickers`, `weights`, `beta`, `mu_capm`, `sigma_hist`, `sigma_capm`
+**Example row**
+tickers: "VOO,QQQ,IBIT"
+weights: "0.53,0.41,0.06"
+beta: 1.18
+mu_capm: 0.153
+sigma_hist: 0.209
+sigma_capm: 0.191
 
-If you publish the dataset to Hugging Face, set the repo to:
-`your-username/LensIQ-Synthetic-Portfolio-1000` and include this README.
-
----
+markdown
+Kopiér
+Rediger
 
-## Project Structure
+## Files
 
-    .
-    ├─ app.py
-    ├─ Dockerfile
-    ├─ requirements.txt
-    ├─ data/                 # generated CSVs
-    └─ README.md
+- `investor_profiles_<timestamp>.csv` — 1,000 synthetic candidates.
 
-## License
+## Notes
 
-- Code: MIT (or your choice)  
-- Dataset: CC BY 4.0
+- Requires the user to include `VOO` in **U** (market proxy).
+- Returns/σ are computed from monthly data; figures are annualized.
+- For reproducibility, a fixed RNG seed is used in the app when sampling mixes.