Data

The Alvíss AI application has a set of strict requirements on the types of data it can work with. This section will run you through the file formats, conventions, and assumptions that this application is operating under. The application supports models, data, and insights on a granular level.

• Alvíss AI is a flexible platform that allows you to add various data sources you need for your specific business.
• We support Sales, Price, Media, Customer Experience, Brand, Visits, Events, Macro, Competitor, and Weather data out of the box.
• Each of these data sources needs to adhere to a specification in order for the platform to understand them.
• While it’s possible and recommended to add all the data sources you can get your hands on, the only requirement is “Sales”. This means you can start out simple and move to a more complex setup later.

Data explanation

Alvíss AI needs a number of data sources to function properly. The main data sources supported are:

• Sales: Core dataset tracking units sold, price per unit, and optionally profit per unit to model business outcomes like revenue and demand.
• Brand: Metrics on your brand's market perception, such as awareness, liking, penetration, and customer experience tracking.
• Competitor Brand: Similar perception metrics for competitors' brands to analyze market dynamics and competitive effects.
• Distribution: Metrics on your distribution network, like number of stores, sales agents, or offices, to model operational presence.
• Competitor Distribution: Distribution metrics for competitors to understand their market reach and impact on your performance.
• Media: Investments and metrics (e.g., impressions, clicks) for your advertising campaigns across channels like display or Google.
• Competitor Media: Media investments by competitors to model competitive advertising pressure.
• Competitor Price: Pricing data for competitors' products to assess price elasticity and market competition.
• Customer Experience: Investments in CX improvements (e.g., IT, human resources) to estimate ROI on customer satisfaction efforts.
• Events: Time-bound occurrences like outages or strikes to account for disruptions in modeling.
• Extra: Custom variables that don't fit other categories, for flexible inclusion of unique business metrics.
• Macro: Macroeconomic indicators like unemployment or consumer confidence to capture broader economic influences.
• Visits: Traffic metrics to websites or stores from channels like direct or paid search, to link engagement to conversions.
• Weather: Meteorological data like precipitation or wind to model environmental impacts on weather-sensitive operations.

These data sources have a specific [File Format](./Format/File Format.md) that needs to be adhered to in order for the application to understand them.

The data sources need to have a coherent data [Periodicity](./Format/Periodicity.md) meaning that the values need to be observed and registered in the same interval.