Efficient route planning with Data

Interflora Finland
A great deal of Interfloras cost structure consists of the payments related to delivering goods. A deeper analysis of current logistics and a framework for a new model was hence needed to improve on the current model and increase the company profits. In today's business climate, delivery logistics encompass a major cost for many businesses and the companies that succeed put a fair amount of smart thinking into devising a payment framework.


Orders analysed


Kilometers travelled


Hours of transport in 2016


Possible yearly savings


Finding an optimal route planning model for delivering goods and their related payments while simultaneously keeping the end customer, florist members and head quarters satisfied. Every party wants to save money while raising the bar on service level. Routing of orders to correct shop based on variables provided by business while making sure the same day deliveries are fulfilled consistently.


Distrikt13 set out to analyse more than 5 years of Interflora shopping and delivery data to devise a data-driven intelligent logistics framework that would propel Interflora into the 21st century alongside the aforementioned businesses. First we cleaned up the provided data and added some geographical variables, e.g. salaries, petrol costs and national holidays. With the help of GoogleMaps and HERE Maps API's we used the address data to retrieve coordinates for delivery destinations. Conducting interviews and quantitative polls with the florist shop owners to shed some light on daily routines. After the data was prepared for modelling and interviews were conducted, we researched how to current model was performing and lastly moved on to provide payment and order routing models that would significantly reduce the cost per delivery in comparison with the current model.

Tools of choice
Google Cloud Platform
Google GeoLocation API
Scikit learn
Pandas PyData
Jupyter notebook

The insights revealed have paved new ways for our business to develop.

Tommy Nyman

Head of Digital, Interflora Finland


We provided three models based on our insights that could be applied standalone or in combination. The models and their affects were provdided as a Jupyter notebook where one can see the calculations and its explanation. Interflora evaluated the effect of the models through a heatmap with value inputs and instantly saw the changes in the logistics pricing landscape. The board got to evaluate options and decide what is best for the business. To grasp the current situation in birds-eye-view, Distrikt13 built a visual representation tool to highlight delivery hotspots on a map that could be adjusted by time, this was an important asset to display needs of drivers on peak seasons vs. daily business. As a cherry on top, a Proof-of-Concept router for the driver to calculate optimal route to drive based on current situation in real time (orders left, destinations, desired time of deliveries).

Want to know more?
Kaarlo Niinistö
Kaarlo Niinistö



Related Work

Modern eCommerce platform

Interfloras service is highly valued by customers during holiday seasons. By examining peaks in load we noticed that queries increased approximately 63 times in relation to the normal load while writes needed capacity for 12 times more. More than a third of Intefloras customers want to have their greetings delivered during the same day and cutoff time for service regions differ. Customer data ownership was at a third party service provider, making sense of the data was a communicational burden rather than technological. System development had stagndated due to will of paying for it since IP's would remain at the provider and higher customer expectations created challenges for the way of doing business today.

Read more

Campaigning for Valentine's

Valentines Day is one of the biggest business opportunities for Interflora and they were looking to hit their sales targets. With a lot of competition in the gift sector for Valentines, Inteflora had to find a way to stand out as memorable option. Customer segments wanted to be reached with unique messages while keeping the media budget to a minimum.

Read more