Operational issues in cloud kitchen

Abstract

Bengaluru is home to restaurants from all over the world. We can find all types of cuisines in this place, from the United States to Japan, Russia to Antarctica. Bengaluru has everything one could possibly want delivery, dine-in, pubs, bars, drinks, buffets, and desserts. With recent evolution of tech-based startups like Zomato & Swiggy and more cuisine focusing delivery chains, every day, there are more restaurants opening. By FY20, total number of eateries' number stood at 12,000 eateries yet without the market reaching its saturation. New eateries are appearing every day and find it challenging to compete with restaurants that have already achieved success. The main problems that they continue to face include high real estate prices, rising food prices, lack of qualified workers, a disjointed supply chain, and over-licensing. Over the top of it, Covid-19 crisis has disrupted the market in two ways - first, new chains want to invest less initially, and focus on more delivery or takeway counters, referred to as cloud kitchen & second, consumers want to eat at the comfort of their homes avoiding unnecessary travel or use the convenience while working from home. This research intends to analyse the area's demography and culinary culture to help new restaurants, specially cloud kitchens to make an informed decision where to open new branches. In addition, we also tried to find out key differences of cloud kitchen versus other restaurant types. The most significant benefit is that it will assist new restaurants in selecting their theme, menus, cuisine, price, etc. for a certain area. It also looks for culinary similarities among Bengaluru neighbourhoods. People will be able to select a restaurant based on the analysis and several other criteria. We have tried to identify the elements that might influence tfie launching of a new restaurant in a community with these questions in mind. Reviews for each restaurant are also included in the dataset, which will aid in determining the location's overall rating. To achieve the above, we have used Python as analysing platform to process data, create meaningful graphs (screenshot attached at relevant places) and come up with related insights. Link to the python notebook has been provided in the references section. Basic data processing along with linking addresses to geo locations have been carried out. In the end, we have also performed clustering and hypothesis testing to support our arguments, all performed in the same python code link.