Python in supply chain management
After 95 years of its operations, in 2016, it was the first time when Qantas saw a handsome profit of $1.5 billion and that was after facing a historical loss of $2.8 billion just two years before. So, what exactly made Qantas turnaround?
Over a period of years, Qantas observed that a major problem faced by them was their inability to come up with accurate demand forecasts for inventory planning. This was due to the fact that they were using a very old demand planning and forecasting system. Besides that, another problem they faced were due to heavy expenditure that would go towards maintaining 28,000 components and 60,000 expendables. Qantas lacked visibility of systems and data generated in the organization. It lacked efficiency.
After several boardroom meetings and consultation, Qantas decided to utilize the power of Internet of Things. It was decided that IoT would help them in connecting different parts and predictive analytics to generate insights from them. Also, an advanced forecasting algorithm would be used for service parts inventory planning. This IoT initiative reduced the inventory costs of Qantas by 20% and improved its service efficiency. The company could then predict the possibilities of failure of any aircraft parts and do the replacement or correction before the problem could escalate and affect the company operations.
Just like Qantas, there are many stories of data science analytics helping solve major supply chain challenges in multiple industries. A fortune 100 Consumer Packaged Goods company was packaging and shipping huge volumes of products across the globe. For them selecting just the right carriers and routes to optimize logistics was a big challenge. By making use of data analytics in logistical operations, the company came up performance metrics. This metrics helped them to evaluate different carriers for cost effectiveness, driver performance, lead times, and other such parameters. The result was a saving of $1 million in every category of transportation and an enhanced warehouse planning.
Machine learning, another major discipline within the umbrella of data science has transformed supply chain tremendously. Here are a few things that ML could do to bring transformation:
- Improved efficiency and capability of handling large data sets, accuracy of demand forecasting is significantly improved.
- Collaborative real time supply chain networks are created, freight costs are reduced and delivery performances are improved.
- Visual pattern recognition helping manufacturing organizations perform inspection and maintenance work remotely.
Examples listed here are some of the ways Machine Learning could help supply chain, but there are many such more examples yet to come. A supply chain system is a vast system of connected processes that influence each other contributing to the effectiveness of the final offering that goes to the market. A typical supply chain begins with raw material procurements from suppliers after which the next stage is of manufacturing, warehousing, and distribution. Data science programming can be used at every stage of a supply chain network.
ProcurementProcurement is one important part of any supply chain. The efficiency of procurement affects the entire supply chain and is a major factor in determining the market price of a product. With the use of data science powered digital supply chains, procurement strategies can be drastically improved. Artificial Intelligence can help predict demand and sources for innovation. Automation algorithms can be implemented to sense material demand, categorize expenditures, conduct supplier audits, and report procurement performance.
ManufacturingWhen lean principles became popular, manufacturers started to focus on cutting costs to increase efficiencies by reducing wastes. With data science comes the next stage of manufacturing process evolution. The goal changes from lowering costs to creating disruption. Data science solutions like big data, analytics, industrial internet of things, and robotics can help transform factories into smart-factories. From regulating inventory to maintaining equipment’s to handling products, the entire production process can be chained into a seamless flow with automation increasing efficiency of these systems. As per a Gartner report, with the use of manufacturing analytics, 88% of the companies using data science in manufacturing were able to improve responsiveness of their systems. 81% saw a clear improvement in their capacity utilization and 75% were able to take decisions faster. Predictive maintenance, a concept that has emerged only recently, can help companies prevent equipment breakdowns and reduce costs using the power of insights gathered from IOT assets.
WarehousingThe world is going to see new kinds of warehouses that are more agile, hyper-connected, and innovative. In these smart warehouses, technology would take care of most processes including receipt of manufactured products, sorting, organizing, picking, and packaging for distribution. Warehousing Robots can help automate the process of picking and brings products from stored locations to the points of packaging. With AI further supporting such robots, a program can provide an optimum pickup route in large warehouses and help packers determine the most efficient way of space utilization for packaging. RFID scanners would communicate with the internal warehouse management system automatically giving information on which products are picked, packed, sorted, loaded or transported.
DistributionOnce a product is ready, it has to reach the target market through a distribution channel. Product needs to be carefully packaged, safely transported and supplied to the distributors, wholesalers, retailers or to customers directly. In either case, logistics optimization remains high on priority. Toughest challenges in logistics today can be solved using data science programming. An example of this is logistics network optimization of Mahindra & Mahindra tractor division. With the use of linear optimization models, M&M was able to identify optimum locations considering handling and freight costs. The company used Python programming to develop scenarios for identification of best solution for transportation of its products.
Data science creates smart supply chains that are cost-effective, efficient, accurate, and optimized over traditional supply-chain systems. It is going to create a smart world with smart supply chains that can operate automatically and communicate with each other throughout the supply chain network. In the near future, companies would know when, where, and how of every activity in their supply chain network.
References
https://www.ptc.com/en/service-software-blog/IoT-predictive-analytics-service-parts-savings
https://www.rtinsights.com/qantas-airlines-takes-off-with-real-time-supply-chain/
https://fractalanalytics.com/casestudies/supply-chain-management-iidentify-the-most-efficient-shipping-carriers-through-advanced-analytics/
https://www.forbes.com/sites/louiscolumbus/2018/06/11/10-ways-machine-learning-is-revolutionizing-supply-chain-management/#2a927d23e370
https://strategywise.com/8-uses-of-data-science-in-manufacturing/
https://www.analyticsindiamag.com/mahindra-mahindra-solving-logistics-network-optimization-challenge-data-analytics/
https://selecthub.com/warehouse-management/smart-warehouse-systems/
