# Pandas Time Series Analysis - Part 3: Holidays(Nepali Context)

Hey friends, namaste! If you’re diving into time series analysis with pandas and need to handle Nepali stock market schedules, you’ve come to the right place. In Nepal, the stock market (e.g., NEPSE) typically operates from **Sunday to Thursday**, with **Friday and Saturday** as weekends. Plus, there are unique national festivals and holidays—like Dashain and Tihar—that might close the market as well.

Let’s walk through how to configure pandas so your date ranges properly exclude non-trading days and incorporate any special holidays.

---

## 1\. Setting Up Your Nepali Stock Data

### Scenario

* You have a CSV file of stock price data from **1st July** to **21st July**, but it’s missing a Date column.
    
* The NEPSE runs **Sunday to Thursday**. That means **Friday** and **Saturday** are not trading days.
    
* You want your time series to skip those days automatically.
    
* You might also want to exclude specific holidays (e.g., major festival dates).
    

### Basic Imports

```python
import pandas as pd
from pandas.tseries.offsets import CustomBusinessDay
```

Assume you already have your CSV file in a DataFrame:

```python
df = pd.read_csv('nepse_stock_data_no_dates.csv')  # Contains prices but no dates
```

---

## 2\. Defining a Custom Business Day for Nepal

In Nepal, the stock exchange is typically closed on **Friday and Saturday**. Pandas’ `CustomBusinessDay` can handle this through the `weekmask` parameter.

```python
# In Nepal, workdays: Sun=1, Mon=2, Tue=3, Wed=4, Thu=5
# Weekend: Fri (6) and Sat (7) are off.
nepal_bd = CustomBusinessDay(weekmask='Sun Mon Tue Wed Thu')

# Let's generate a DateTime index from July 1st to July 21st 
date_index = pd.date_range(start='2023-07-01', end='2023-07-21', freq=nepal_bd)

print(date_index)
```

When you print this out, you’ll notice:

* It **skips** every Friday and Saturday automatically.
    
* Only includes Sunday through Thursday.
    

Now, assign this custom index to your DataFrame:

```python
df.index = date_index
print(df.head())
```

Your DataFrame’s rows should match the actual trading days, no more “weekend” gaps.

---

## 3\. Handling Nepali Holidays and Festivals

Of course, NEPSE isn’t only closed on weekends. Major festivals like **Dashain** or **Tihar** can also shut down trading for a few days. If you know those specific dates, you can exclude them by passing a `holidays` list into `CustomBusinessDay`.

### Example: Excluding Dashain & Tihar Dates

Let’s say you know the market will be closed on:

* **24th October 2023** (Dashain)
    
* **12th November 2023** (Tihar)
    

Here’s how you’d define that:

```python
nepal_bd_holidays = CustomBusinessDay(
    weekmask='Sun Mon Tue Wed Thu',
    # List out the exact Gregorian dates of Nepali holidays (format: YYYY-MM-DD)
    holidays=['2023-10-24', '2023-11-12']  
)

date_index_festivals = pd.date_range(
    start='2023-10-01',
    end='2023-11-30',
    freq=nepal_bd_holidays
)

print(date_index_festivals)
```

Check the output to confirm that **October 24th** and **November 12th** are missing. Perfect for aligning your data with actual NEPSE trading days.

---

## 4\. Creating a Custom Holiday Calendar

If you want a more robust or dynamic holiday system (e.g., changing festival dates each year), you can subclass `AbstractHolidayCalendar`. This might be handy if you’re working on a long-term project and want to keep a central record of all the major holidays.

```python
from pandas.tseries.holiday import AbstractHolidayCalendar, Holiday

class NepaliHolidayCalendar(AbstractHolidayCalendar):
    """
    Nepali holiday calendar for major festivals. 
    Just an example – fill in the actual dates/rules as needed.
    """
    # Example: Dashain (10/24), Tihar (11/12)
    rules = [
        Holiday('Dashain', month=10, day=24),
        Holiday('Tihar', month=11, day=12),
    ]

# Create a CustomBusinessDay with your new holiday calendar
from pandas.tseries.offsets import CustomBusinessDay

nepali_calendar = NepaliHolidayCalendar()
nepal_bd_calendar = CustomBusinessDay(weekmask='Sun Mon Tue Wed Thu', calendar=nepali_calendar)

# Generate a date range for Oct-Nov
date_range_calendar = pd.date_range(
    start='2023-10-01',
    end='2023-11-30',
    freq=nepal_bd_calendar
)

print(date_range_calendar)
```

This approach is more scalable if you want to keep adding new holiday rules. Just update the `NepaliHolidayCalendar`class, and everything else will follow.

---

## 5\. Putting It All Together

Let’s say you have a CSV with trading prices from July 1st to July 31st, but no date column. You know:

* NEPSE trades **Sunday to Thursday**.
    
* You want to skip weekends (Fri, Sat).
    
* You also want to skip July 9th as a special closure day.
    

Here’s how you can do it in one go:

```python
import pandas as pd
from pandas.tseries.offsets import CustomBusinessDay

# Step 1: Read the CSV
df = pd.read_csv('nepse_stock_data_no_dates.csv')

# Step 2: Define a custom business day
nepal_bd_custom = CustomBusinessDay(
    weekmask='Sun Mon Tue Wed Thu',
    holidays=['2023-07-09']  # special closure day
)

# Step 3: Generate a date range
date_index_custom = pd.date_range(
    start='2023-07-01',
    end='2023-07-31',
    freq=nepal_bd_custom
)

# Step 4: Assign this index to your DataFrame
df.index = date_index_custom

print(df)
```

Voilà! Your DataFrame now lines up exactly with NEPSE trading days for July, minus that special holiday.

---

## 6\. Why This Matters

* **Accuracy in Analysis**: Financial, forecasting, or machine learning models that use actual trading days can be more accurate.
    
* **Planning & Scheduling**: If you’re building a tool for local investors or analysts, they’ll appreciate correct date alignment for NEPSE.
    
* **Global Comparisons**: If you compare multiple markets (e.g., NEPSE with NYSE), each has its own weekends and holidays. Pandas can handle all of it!
    

By fine-tuning your date ranges, you’ll avoid the confusion of “phantom” Fridays or Saturdays when the market’s closed. You can also set up your code to skip major festivals so your data perfectly reflects real-world conditions.

---

## 7\. Key Takeaways

1. **Use** `CustomBusinessDay` to define working days and skip weekends specific to Nepal’s schedule (Sunday–Thursday).
    
2. **Add a** `holidays` list or subclass `AbstractHolidayCalendar` to handle big festival closures like Dashain, Tihar, or any custom off days.
    
3. **Observance rules** can shift a holiday if it falls on a weekend—handy if your system tracks floating holiday policies.
    
4. **Experiment!** The best way to learn is by trying these code snippets in a Jupyter Notebook or any Python environment.