Accessible Big
Connected Data

Three ways cloud services help get more knowledge out of data.

Presented at Fruitnet World of Fresh Ideas — Berlin, 7th Feb 2017

Read the talk
Fresh produce Fresh produce

Every time a strawberry gets picked, or a ship leaves the port, new data is created.

Governments all over the world are doing an amazing job at promoting transparency. Massive datasets are now available on services like Quandl that are accessible with a single line of code.

But having all this information is just the first step. The challenge is to extract knowledge from data, and there are three key ways in which cloud services can help.

1

Upgrade to the Speed of Cloud

Your data grows exponentially faster than your business, your data tools should evolve as well.

As your business scales, the amount of data you generate scales exponentially. Can a cloud service really be faster than analysing data on a computer at your desk? Watch as we demonstrate the power of cloud services to answer a businesses question: How were sales last month?

Cloud companies like us and many others work constantly to build and improve the technologies that deliver this kind of speed – indexing, caching, predictive loading, efficiency optimisations and in-memory processing.

2

Hire a Virtual Analyst

Sometimes it's easier to just ask someone.

You're probably familiar with virtual assistants like Siri, Cortana and Alexa. As well as playing songs and ordering washing powder, the technology behind them offers great opportunities for business-facing services. Cortana can be integrated with BI tools, and Amazon has opened up Lex – the brains inside Alexa – for developers to use in any application.

What that means to you is that specialized tools can now be offered at very low cost. Here's an example of Fresh4cast's virtual analyst Saga helping answer a question directly: Why are Melon stocks high?

3

Extract Gold from your Data with Forecasting

There are forecasting solutions for every budget

In classic forecasting, several years of data can be used to predict the following few years. Here's a basic example. The confidence interval is shown in blue, and is quite large.

Forecasting Example Graph fit_arima <- auto.arima(pineapple) plot(forecast(fit_arima, 36), main="Pineapple exports")

This forecast only looks at seasonality and long-term trend. It only takes two lines of code, and you can do similar basic forecasting easily in Excel. More accuracy requires more complex models, and cutting-edge forecasting is now using machine learning techniques.

Using powerful bespoke forecasting, Fresh4cast helped one customer achieve a 35% error reduction compared to their manual forecast for Strawberry yield in the UK, resulting in better workforce planning, better sales planning, and less waste.

35% reduction in mean average percentage error

Fresh4cast's Live Forecasting Tool

Fresh4cast combines cutting-edge forecasting algorithms with our powerful, bespoke cloud platform to provide real-time, high quality updates.

This is available on the same platform as sales data, our global supply data and your manual forecasts, and can be integrated with our virtual analyst, Saga.

Learn More