By continuing to browse this website, you consent to our use of cookies, as well as to our Terms of Use and Privacy Policy which provide additional information about how we process your data. This website uses cookies to enhance your user experience. Please read our Cookies Policy for more information on how we use cookies, as well as instructions on how to disable cookies. You may disable cookies through your internet browser settings, however this may result in some parts of the website not working properly for you.

INSIGHTS / INTERVIEWS

Raw Data: Bringing new predictability to harvests

Spain · Sep 30, 2019· By Gareth Gardiner Jones

The Raw Data team ©RawData

Spanish ML, big data startup helps farmers perfect wine and fruit production in a fast-growing precision agtech sector

With climate change affecting crop yield and quality, farmers are increasingly looking to technology – particularly predictive data – for solutions. Raw Data is a young Barcelona-based startup running an ML- and big data-based predictive SaaS platform for stone fruit farmers and winemakers. 

“Our objective is to try to reduce the uncertainty of what will happen in cultivation by identifying via predictive data the possible challenges affecting each crop," Albert Duaigues, CEO and co-founder, of Raw Data said. "For example, in viticulture, we have created a predictive model for quality control and to evaluate the maturity of the grape to define the most precise dates for harvest.

“The sector is aware of climate fluctuations, making the needs more palpable for our type of service. Predictive algorithms have to be stronger, too, with ever more data input required," Duaigues told CompassList at this year's Smart Agrifood Summit in Malaga.

Precise predictions provide savings of up to 35% in personnel in the fields, as well as taking preventative action on diseases and adverse weather, he said.

As a major global agricultural producer, Spain wants to be at the forefront of 4.0 innovation in agtech. Despite relatively low levels of digitalization in the agricultural sector worldwide, change is underway with predictive data that can help to boost production and reduce resource-use. Indeed, the global smart agriculture market is forecast to reach US$15.34bn by 2025, based on CAGR of 13.09% from 2017.

Complementary skills 

Duaigues was already an entrepreneur and had worked as a marketing, sales and communications manager when he returned to his studies in 2017 at Barcelona's International University of Catalonia (UIC). There, he undertook an Executive Master's in Big Data Science and focused his final project on solving problems in the agricultural sector. He also got together with data scientist and engineer David Olmo, whose experience in ML proved crucial in the creation of Raw Data. They finished their platform in January 2019 after nine months of development.

Complementing Duaigues' commercial and management experience were also full-stack developer Alejandro Martin and marketing and finance expert Emma Sabater Font, who complete the current Raw Data team. The company has been based from the outset at Spanish ITC giant Telefonica's Barcelona accelerator, which has helped the co-founders fine-tune their predictive models and bring Raw Data to its current MVP stage.

The company has found a more open marketplace by selling its services mainly to agricultural conglomerates, co-operatives, agri technicians and consultants, rather than to individual producers.

“The wine sector especially is more open to innovation and quick to take decisions," said Duaigues. "At this level, there is already a lot of pooled data from producers, strengthening the accuracy of predictive algorithms.”

Raw Data requires as little as two years' data to produce accurate predictions, such as the size of cultivation and optimum harvest-ready date.

Leveraging agtech experience

Duaigues' experience founding and managing his previous startup, SaaS eFoodPrint Services (established in 2013 and still in operation under his management) also gave the team a head-start in the sector. The company provides agro producers with a sustainability assessment by using the producers' agronomical data to define its carbon and water measurement.

Duaigues developed his software for that company with his MBA research group at Rovira i Virgili University in Tarragona in 2010. The business has helped to develop commercial relationships in the agro sector and even to cross-sell Raw Data to a couple of customers. “However, we see higher growth possibilities for Raw Data as environmental sustainability is a corporate plus not an agricultural producer's mission," he said. "Raw Data is helping them in their core business, i.e., production.” 

Raw Data already has several paying customers prior to the formal business launch planned after seed investment has been received. These include one of Catalonia's main co-operative fruit producers, for whom Raw Data predicts production volumes and harvesting. For another client, a major cherry producer, Raw Data's predictions of harvest volumes have allowed the company to plan exact volumes for its clients well in advance. 

Though the startup is not the first to offer ML- and big data-based agtech insights, even within Spain, “each [agtech company] is specialized in a different crop and area," according to Duaigues. "Our fields of expertise are stone fruit and wine growing, disease control and harvesting analysis.

"Because agriculture is very complicated, with many variables affecting the development of crops, there is potential for all predictive services because they empower producers and analysts to take advance and better-informed decisions than previously possible and to make their businesses more profitable.”

Trial and no error

Raw Data's ML algorithms are built on the foundation of a wealth of data input, including on production parameters, meteorological data, ripening controls, planting density and soil type. In its disease control module, diseases such as mildew can be detected a week before their appearance using predictive climate models applied to each tract of land.

“The more variables that are provided, the more reliable and competitive models we can create,” said the CEO.

Once a potential client is interested, Raw Data creates a database and builds algorithms with as much historical data as possible, to which agronomic variables such as temperature, soil salinity and radiation are further factored in. The robustness of the algorithm is then tested by applying it to past cultivations to see if the result is validated. If so, the algorithms are proposed to the client in the form of a year-long SaaS access to the platform.

Prices vary according to the crops surveyed and the amount of data entered. Once a customer signs up, their personalized dashboard will then display continually refined results, factoring in new data inputs such as weather conditions as they appear.

Though almost all of the process is digitalized and not conducted in person, Duaigues believes personal contact also plays an important role in this sector, building trust and also permitting the company to better know and cater to its customers. Word-of-mouth recommendation is also a highly important form of marketing in the relatively close-knit community of producers of each crop in each region of Spain.

Raw Data is currently seeking €150,000 for its first seed funding round and has secured a large part of the amount from an unnamed investor. The monies will be used for further technical and commercial development of the platform, including its expansion across Spain and an assessment to select one of its two international investment plans for rollout. The first foresees Europe-wide expansion, covering major wine and fruit producers such as France and Italy, while the other would target the same sectors but in South American agricultural giants such as Argentina and Chile.

Edited by Matt Stanley