RawData Predictions
Harvest forecasting module of the Spanish farm platform RawData. It predicts crop volume in kilos and ripening dates fifteen days ahead, combining growers' own historical records with 447 weather station and satellite factors.
What is RawData Predictions?
RawData Predictions is the harvest forecasting module of RawData, the farm management platform published by the Spanish company RAW DATA SL. It answers a single question that governs a whole season: how much will come in, and when will it be ready. The module forecasts two things. The first is harvest volume in kilos, so a producer can plan sales, decide when to ship, and avoid flooding the market at the wrong moment. The second is the ripening date, predicted fifteen days ahead, together with sugar degree, pH and acidity, so the picking window can be chosen on evidence rather than instinct.
The method runs in three stages. RawData first recovers the customer's own historical agronomic records, in whatever format they exist, and enriches them with 447 additional factors drawn from weather stations and satellites. It then applies machine learning and, crucially, validates the resulting model and demonstrates its reliability to the customer before anything goes live. Only then are platform access and user accounts opened. The published headline is 95% reliability, with the model described as validated across 32,606 plots.
The forecasts are consulted inside Agrodata, RawData's mobile app for technical crop management, which also carries the field notebook, SigPac maps and visit logging, and which works in areas without network coverage. Alongside the two forecasts, the module records field visits — yield counts, pests, diseases, control points — and lets a technician replay the history of any plot at a chosen date.
The most detailed published case is Covides, Catalonia's leading first-degree wine cooperative: 2,000 hectares, 650 growers, 20 million kilos in 2021 and 8 million bottles. Over two campaigns its ripening error fell from the 10–25% typical of traditional methods to 5–7% at fifteen days, roughly tripling forecast precision. Across the whole platform the publisher claims more than 200 client companies. Support is delivered around the clock by agronomists rather than a call centre, and the offer carries a two-month money-back guarantee.
What it does
- Forecast harvest volume in kilos, plot by plot and variety by variety
- Predict the ripening date fifteen days ahead, with sugar degree, pH and acidity
- Log field visits, including yield counts, pests, diseases and control points
- Centralise agronomic and commercial information into a single 360-degree view
- Replay the history of any plot at a chosen date
- Feed harvest figures into an ERP in real time
When to use RawData Predictions / When not to
A quick filter to help you decide if RawData Predictions is the right fit.
When to use RawData Predictions
- Cooperatives that must sequence a large harvest intake, as in the Covides case of 20 million kilos across 650 growers and 2,000 hectares
- Technical directors who plan the harvesting window and the commercial calendar together
- Farming businesses that already run an ERP and want field data to reach it without double entry
- Producers holding several seasons of agronomic history, which is the raw material the model needs
- Operations trying to cut the travel and staffing cost of manual ripening checks
When not to use RawData Predictions
- New operations with no historical records to hand over, since the method starts by recovering past agrodata
- Buyers looking for a free or commitment-free tool, as the platform requires a minimum one-year subscription
- Anyone outside crop farming, because the product is built entirely around agronomic use cases
- Teams wanting to buy the prediction module on its own, as no standalone price exists for it
- Users who need an interface in a language other than Spanish, Catalan or English
How to use RawData Predictions
A typical end-to-end flow, from setup to results.
- Book a free demo with an agronomist through the product page, with no commitment attached
- Hand over your historical agronomic records, in whatever format they currently exist
- Let RawData enrich them with the 447 weather station and satellite factors it adds
- Review the reliability that RawData demonstrates for your model before it goes into service
- Take delivery of platform access and create the user accounts your team needs
- Go through the guided rollout, roughly a month covering data loading, training and support
- Consult volume and ripening forecasts day to day in the Agrodata mobile app, including in areas without coverage
- Record field visits, counts and control points from mobile so the history keeps building
- Connect your ERP through the authenticated public API if you want figures to flow automatically
Pros & Cons
Pros
- The reliability claim is backed by a named, quantified customer case rather than a testimonial alone
- The model is validated and its accuracy shown to the customer before it is put into service
- No IoT hardware to install, since the forecast is built from existing records, stations and satellites
- Support is staffed by agronomists on a 24/7 basis rather than a generic call centre
- A two-month money-back guarantee limits the risk of committing
- Real-time bidirectional ERP integration through a public API
- The field app remains usable in plots with no network coverage
Cons
- The prediction module has no price of its own, and the pricing page never lists it
- The platform requires a minimum one-year commitment, with displayed prices assuming annual billing
- The 95% reliability figure comes with no published methodology or measurement scope
- The API documentation link points to a dead anchor, so no technical reference is publicly readable
- No data processing agreement, no subprocessor list and no declared hosting location
- No security certification such as ISO 27001 or SOC 2 is claimed anywhere on the site
- Strong Spanish orientation, from SigPac and MAGRAMA references down to a three-language site
Pricing & Plans
There is no permanent free plan. RawData does not publish a price for the Predictions module itself; the pricing page lists only the two base modules, of which the relevant one here is Agrodata at 49 EUR per month, the environment that carries the prediction features. That figure assumes annual billing and a minimum one-year subscription, and the final amount varies with the number of plots or active workers. A free demonstration and a two-month money-back guarantee are offered instead of a free tier.
- limited users
- activity logging
- field notebook and Global G.A.P
- integration with other platforms
- offline connection
- cost reports
- MAPAMA/SIEX autocompletion
- unlimited users
- biometric and NFC time tracking
- absences
- digital work reports
- worker geolocation (a separate product line
- outside the scope of this record)
- Agronomics
- Mechanics
- Productividad
- Portal del empleado
- Prenomina
- Economia
- Portal Socios
- Sostenibilidad
- The two base modules can be combined into an all-in-one configuration
Data, GDPR & hosting
A consolidated view of how RawData Predictions handles your data.
GDPR overview
The privacy policy anchors processing explicitly in Regulation (EU) 2016/679 and names consent under Article 6.1(a) as its legal basis. It lists the full set of data subject rights, including access, rectification, erasure, restriction, objection, portability and the right not to be subject to purely automated decisions. Rights are exercised by writing to lopd@agrawdata.com or by post to the Mollet Val address, and the Spanish supervisory authority is named with its address and phone number. Consent can be withdrawn at any time. Concrete gaps remain: no data processing agreement is published, no subprocessor list exists, no security certification is claimed, and the policy carries no effective date or revision date. As a Spanish company, the publisher sits inside the Union and needs no Article 27 representative.
Who owns the data?
The data controller is RAW DATA SL, a Spanish company registered under tax number B67275883. Its privacy policy gives a registered office at Paseo Ronda 9, floor 1, door 2, 08100 Mollet Val (Barcelona), while the legal notice states Rambla Mercedes 26, 08024 Barcelona. Users remain responsible for the accuracy of what they submit and must report any change. Clause 10 states that personal data will not be transferred to third-party companies or national entities without the user's prior consent. Access is restricted to authorised staff, and the policy describes no subprocessing arrangement and no international transfer.
Reuse rights
The privacy policy lists the categories collected: identification details, contact details, banking details needed to process payments, and device location captured at clock-in time. Location is optional, requires the user to switch on the permission explicitly, and can be revoked at any time without affecting other functions. The stated purposes are limited to delivering the website and app services, handling requests and incidents, taking payment, and sending commercial communications, all on the legal basis of consent under Article 6.1(a) GDPR. On the product side, RawData states that it builds each prediction model from the customer's own historical agronomic records, enriched with 447 station and satellite factors. Nothing in the documents grants the vendor a broader reuse right, and nothing describes customer data being pooled to train a shared model.
Data retention & training
Hosting summary
RawData publishes no information about where customer data is hosted. No country and no region are named anywhere on the site, and no infrastructure provider is identified, whether AWS, Azure or Google Cloud. The resolved IP address for the domain belongs to a Cloudflare anycast node and therefore says nothing about where data is actually stored. What can be established is jurisdictional rather than technical: the publisher, RAW DATA SL, is established in Spain and therefore inside the European Union, and its privacy policy states that processing follows Regulation (EU) 2016/679. The policy describes no international transfer and no transfer mechanism such as standard contractual clauses, which is consistent with a purely domestic operation but is never stated as such. Buyers with a hosting requirement should ask for it in writing, since nothing on the site answers the question.
Things to keep in mind
Risks and trade-offs to weigh before adopting RawData Predictions.
- Treating a forecast as a certainty: the 95% reliability figure has no published methodology or scope, and a season can still break the pattern
- Letting predictions replace field judgement, when the publisher's own reference customer stresses that technicians and their visits remain necessary
- Committing for a minimum year on a module whose price is never listed separately, on top of a base subscription that scales with plots
- Handing over several seasons of agronomic history without settling contractually what the vendor may do with it, since the model is built from that data
- Relying on a compliance file that lacks a processing agreement, a subprocessor list and any declared hosting country
- Assuming the published address is settled, when the legal notice and the privacy policy give two different registered offices
- Planning an ERP integration on an API whose public documentation link leads nowhere
Setup & Integrations
Technical difficulty
Low for the customer, because the vendor does the technical work. RawData recovers the historical records, builds and validates the model, then opens platform access and user accounts. The guided rollout runs about a month and covers data loading, team training and hands-on support, with the publisher claiming it needs no involvement from an internal IT team. The one real prerequisite is having usable agronomic history to supply. Connecting an ERP is the harder part, since it goes through an authenticated API whose documentation is not publicly accessible.
Deployment
Integrations
Supported languages
Behind RawData Predictions
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What exactly does RawData Predictions forecast?
How reliable are the forecasts?
Do I need to install sensors or IoT equipment?
How much does the Predictions module cost?
Is there a free plan?
How long does implementation take?
Can I connect the forecasts to my ERP?
Does the app work in fields without network coverage?
Which languages is the service available in?
Who publishes the tool?
Should you pick RawData Predictions?
RawData Predictions is a narrow tool that does one hard thing and documents it better than most. Forecasting how many kilos will arrive and when the fruit will be ready is exactly the decision that governs staffing, logistics and price, and the published Covides case gives the claim substance: ripening error cut from a range of 10–25% down to 5–7% at fifteen days, across a cooperative handling 20 million kilos. Two design choices deserve credit. The model is validated and its accuracy shown to the customer before going live, which is rarer than it should be. And nothing has to be installed in the field, because the forecast is built from records the grower already holds, enriched with station and satellite data.
The reservations are mostly about what is not published. There is no price for the module itself, so the practical entry point is the 49 EUR per month Agrodata base that carries it, on a minimum one-year commitment. The 95% reliability headline arrives without a methodology. The API documentation link is dead, and the compliance file is thin: no processing agreement, no subprocessor list, no stated hosting country, no security certification. The registered address even differs between the legal notice and the privacy policy.
It suits an established producer or cooperative that already keeps several seasons of agronomic history and runs an ERP worth feeding. It suits Spanish operations best, given the SigPac and MAGRAMA groundwork. Anyone starting from scratch, or unwilling to commit for a year, should look elsewhere. Ask for the reliability figures on your own crop during the demo, and settle the data questions contractually before signing.
- Choosing a selection results in a full page refresh.
- Opens in a new window.