ANLY482 AY2017-18T2 Group10 Analysis & Findings
Due to confidentiality, we will not be able to upload any charts onto the wiki. The fully disclosed analysis report is available in our Interim Report submission.
EDA For Inventory Data
80/20 Analysis
Given that the sponsor has a few hundred items in their inventory listing, we would like to identify which items make up a larger percentage of the total usage cost. We could then choose to focus on items which make up the bulk of the inventory costs, as the impact of any inventory forecasting model would be greater for those items.
From our analysis, we have found out that the top 70 items make up 60% of the total usage cost for the past two years, while the top 140 items make up almost 80% of the total usage cost for the same duration. The detailed From this we can tell that a substantial proportion of items make up 80% of the total usage value and thus not quite ‘80-20’.
The main cost drivers are largely the meat items, such as different cuts of lamb and beef, as opposed to the vegetables. Seafood items such as crayfish and prawns also take up a high percentage of total usage cost. This is not surprising since consumers want to maximize their “return” i.e. getting the most value for what they pay for. This leads to higher meat consumption since meat is often seen as “higher value” due to its higher cost. Models to be developed will take this into account and cater more accuracy towards predicting these items.
Total Inventory Usage Across Months Breakdown by Outlet
Next, we want to observe the general trend of inventory usage cost over the months to see if there is any seasonal trend in usage cost. Given that there are only two years worth of data, any trends observed might have other confounding factors and may not be completely attributed to the time of year.
Next, we can observe that the December months incurred the larger amount of inventory costs. On the other hand, the month of March for both years incurred one of the lowest inventory costs. Generally, inventory costs are increasing from the month of April up till December. Next, we also want to investigate the proportion of each outlet’s contribution to the total inventory cost by looking at the 100% stacked chart. We can see that the percentage of each outlet’s contribution to the total inventory cost does not vary substantially even though the actual usage values do fluctuate across the months. This shows that each outlet appears to be affected by the seasonality factors by roughly the same amount.
Inventory Usage Across Months Breakdown By Product)
Subsequently, we wish to check if there are any trends present in the inventory usage for individual products across the months. In general, over the entire year, we observe that there is little trend for all the ingredients on the whole and that some ingredients have a more noticeable trend pattern than others. There are, however, some slight spikes in usage for the months of December and August for a majority of the items.
The high usage items (Usage Value > $10,000) appear to have a more noticeable trend than the others. These items include the “XXX” , “YYY” and the “ZZZ” (Yellow). However, we also observe that these items have been discontinued with effect from May 2017. The low usage items (Usage Value < $10,000) on the other hand, show minimal changes across the months and are fairly constant throughout the year. Another observation is on the introduction of new items such as the “YYY- XXX” which was added to the menu only in April 2017 but has seen steady growth.
EDA For Sales Data
---
EDA For PLU Data
---
