​
AI-Powered Demand Forecasting for Travel Technology

​
The Challenge

Travel technology platforms need accurate demand forecasting to optimize inventory, pricing, and revenue management across different markets and seasons. Traditional methods struggle with complex patterns involving seasonality, events, market variations, and changing customer preferences.

​
The Solution: AI-Powered Demand Forecasting

Using Plexe’s AI platform, companies can implement sophisticated demand forecasting without ML expertise. Just describe what you need in natural language, and optionally provide your data - Plexe handles the rest.

​
Quick Start Example

import plexe

# Build the model (with optional data)
model_version = plexe.build(
    goal="""Create a demand forecasting model that predicts hotel room demand 
    for the next 90 days considering seasonality, events, and market segments.""",
    model_name="hotel-demand",
    data_files=["historical_bookings.csv", "events.pdf"]  # Optional - can build without data
)

# Check model status
status = plexe.get_status(
    model_name="hotel-demand", 
    model_version=model_version
)

# Get predictions
forecast = plexe.infer(
    model_name="hotel-demand",
    model_version=model_version,
    input_data={
        "dates": "2024-06-01/2024-06-07",
        "room_type": "deluxe",
        "location": "downtown"
    }
)

​
Using Direct Client

from plexe import PlexeAI

# Initialize with API key (or use PLEXE_API_KEY environment variable)
with PlexeAI(api_key="your-api-key") as client:
    # Option 1: Build with data
    upload_id = client.upload_files(["historical_bookings.csv", "events.pdf"])
    model_version = client.build(
        goal="Create a demand forecasting model that predicts hotel room demand.",
        model_name="hotel-demand",
        upload_id=upload_id
    )
    
    # Option 2: Build without data
    model_version = client.build(
        goal="Create a demand forecasting model that predicts hotel room demand.",
        model_name="hotel-demand"
    )

​
Batch Processing

# Get multiple forecasts efficiently
forecasts = plexe.batch_infer(
    model_name="hotel-demand",
    model_version=model_version,
    inputs=[
        {
            "dates": "2024-06-01/2024-06-07",
            "room_type": "deluxe",
            "location": "downtown"
        },
        {
            "dates": "2024-06-01/2024-06-07",
            "room_type": "suite",
            "location": "resort"
        }
    ]
)

​
Async Support

import asyncio
from fastapi import FastAPI
from plexe import PlexeAI

app = FastAPI()

# Initialize client (can also use environment variable)
client = PlexeAI(api_key="your-api-key")

@app.post("/forecast")
async def get_forecast(data: dict):
    # Using async inference
    forecast = await client.ainfer(
        model_name="hotel-demand",
        model_version="your_model_version",
        input_data=data
    )
    return forecast

# Don't forget to handle cleanup
@app.on_event("shutdown")
async def shutdown_event():
    await client.async_client.aclose()

​
Benefits

  1. One-Line Implementation

    • Create sophisticated models with natural language descriptions
    • Optionally enhance with your own data
    • Instant deployment

  2. Powerful Features

    • Automatic pattern recognition
    • Multi-source data integration (CSV, PDF, images, and more)
    • Real-time predictions
    • Both sync and async APIs

  3. Business Impact

    • Optimized inventory
    • Better pricing decisions
    • Improved revenue
    • Data-driven planning

​
Getting Started

  1. Install the package:

    pip install plexe

  2. Set up your API key either:

    • As environment variable: export PLEXE_API_KEY=your-api-key
    • Or pass directly to the client: PlexeAI(api_key="your-api-key")

  3. Start forecasting!

Transform your travel technology platform with AI-powered demand forecasting today!

QuickstartRecommendations model