TALK TO US
ABOUT YOUR PROJECT

Web Applications That Deliver Real Business Solutions

Form

Form – Python, Libraries (NLP, Keras, Scikit-learn), Machine Learning Frameworks (CNN, RNN)

Client Profile

FORM has been the leading provider of digital forms since 2001. FORM offers the world’s fastest, most accurate, and only integrated task management and image recognition solution for smarter retail execution.

Client Industry

Data Analysis

Business Challenges

The client was in search of sales support software specifically designed for beverage companies. The desired software should be capable of extracting data from menu cards in bars and restaurants, applying it to market research for wines, spirits, cocktails, and other products to achieve target metrics.

Other challenges include:

  • Existing market statistics were based on manually collected data sources.
  • Limited availability of data sources.
  • Difficulty in obtaining information about items listed on the menu cards of restaurants and bars.
  • Processing extensive information from multiple restaurants/bars resulted in;
    • Excessive effort with minimal outcomes.
    • Immense time consumption.
    • Extensive utilization of human resources.
    • Unreliable data.

Solution

  • Plego Technologies developed a beverage sales support software that utilizes machine learning techniques to extract text information from menu cards and automatically segments it.
  • The application processes the data and forwards it to a data warehouse for further analysis by the business intelligence or sales teams.
  • The algorithm in the software enables executives to derive market insights.
  • The application facilitated the expansion of available information pertaining to wines, spirits, cocktails, and their ingredients. This enables beverage companies to make accurate sales and placement decisions.
  • The application is designed with multiple mathematical calculations and logic, employing techniques such as Optical Character Recognition (OCR), Text Planning, and Natural Language Generation (NLG) to retrieve text information from the menu.



  • The text data is further segmented with the help of Deep Learning Artificial Intelligence that includes; the Long Short-Term Memory – LSTM approach. Segmented data is then uniquely categorized using;
    • Convolution Neural Network – CNN to extract data from Images.
    • Recurrent Neural Network – RNN Time Series Data Format.
    • Object Detection Model – ODM, Identify the position of objectives within the Image.
  • The application employs Artificial Intelligence and Machine Learning Statistical tools to explore data and provide more accurate and effective results.
  • Insights from the software help brands identify customer associations with;
    • Liquor.
    • Wines.
    • Beers.
    • Food ingredients.
    • Price.
    • Quantity.
    • and other factors.



  • Developers employed an Open-Source Computer Vision Library to guide and distinguish the menu’s categories, such as;
    • Drinks.
    • Beer.
    • Wine.
    • Meal sections.



  • The team at Plego integrated machine learning technologies to interpret information from menu cards successfully while eliminating manual activities, allowing Businesses to reduce the workforce and time while optimizing accurate results, enhancing the ability to produce and sell the products efficiently.

Technologies Used

Python, Libraries (NLP, Keras, Scikit-learn), Machine Learning Frameworks (CNN, RNN)

Services Provided

Application Development