i have created a model in AWS
contains Sales records by date
for example
Type: Sale,Time:2016-08-01,Success:1 (1 is a boolean)
i want to predict how much Sales will be after 1 month from the latest date (2016-08-01)
which means a combo of Type=Sale AND Time >2016-08-01 and Success=1
any idea how to achieve this
thank u
You need to aggregate your data to a wider array of attributes to be able to use Amazon ML for such predictions. You can use different level of aggregation, for example daily, weekly and monthly.
You should also add any relevant information for the items that you are selling. For example, if you are selling umbrellas, you should add information about the amount of rain on that day, or if you are selling flowers, you should add information about day of the week or proximity to holidays, when people are buying more flowers.
Related
If someone could please advise on how to achieve the following function on my site: I have a business where I ship live animals to customers. The challenge is communicating the weather forecast with the customer, along with the safest day/s to ship for the animals’ wellbeing. This planning is especially key during the peak summer and winter months when temperatures are very extreme. I am looking to automate these functions during the customers checkout process, giving them control of the shipping date. I am looking to implement something like this on my site.
I found another website where I recently placed an order that provided this relevant information for the customer to review before deciding on what day they would like to have the animal shipment shipped out. The table provided a 5 day forecast, which included temperatures for morning and evening at the point of origin, in transit, as well as the customers’ destination address. Any day/s where the temperature is too hot or too cold was marked as unavailable (red) for shipping. Possible shipping dates consisted of two consecutive days where temps are in the safe range and labeled as available (green) Excluding Friday, Saturday, and Sunday. It is also important for the customer to plan ahead and pick a shipping day based on their availability to be home to receive the package the following morning.
I need to calculate the time elapsed between to dates, in days. The values are the creation date and the payment date for every invoice in my model table. I will use this value to classify my invoices: sort by time passed, classyfy them, make a Pareto Chart... etc. So I need to create a calculated column in the inoice model table (can't imagine an approach using measures).
But I have two handicaps:
Tables ARE NOT DIRECTLY RELATED, so I can't use the RELATED function
The invoice can be paid in several installments
So, for each invoice, I need to calculate de time elapsed, in days, between the generation date (in a table) and the MAX payment date of all it's installments (in another indirectly related table)
This:
DaysElapsedInPayment = DATEDIFF(Invoices[InvoiceDate], Max(RELATED(Installments[InstallmentPaymentDate])), DAY)
would work if the tables were directly related an DateDiff whould accept MAX, but no, it's not the case. For your information, if could be usefull to you, this is my invoice-payment model scheme:
Could you help me, please?
EDIT:
As requested, I explain more deeply my data model...
My model looks complex because an invoice can be paid by several payments (part in cash, part in direct debit, for example), so a invoice can have MULTIPLE payments. At the same time, a payment can involve several invoices. Thats why my invoices-payments is many-to-many. And also, a payment (for example a payment by direct debit) can be divided in installments that are monthly sent to bank.
My model has lots of tables, not in english. For my purpose, I mock the needed ones here:
As you can see in the mocked data:
One invoice is paid in a single payment. It's the common case. We only have a single payment date in 'Payments'. That one is the one I need.
Another invoice is paid in TWO payments (part in cash, part payed by VISA), in different days. I need the date of the LAST of both payments.
Last case, I have two invoices that will be paid as a 'SINGLE' payment. Both should have to return the PaymentDate in Payments. But, at the las moment, we agreed with the client to pay them in installments. So installments are created, each one with a DueDate an the date wuere they were finally payed. So I need to get the LAST date of all installments for this payment.
It's complex, I know, but at the end, is about:
Given a value, find related values in anoter table indirectly related
From all values found, get the greatest
The function must perform theese operations, and put the results in a calculated column.
Have you triend creating a Calendar Table with the AUTOCALENDAR function? That way you could create a relation between both tables to make the calculations. Let me know if you need further assistance in order to add this table to the model
I am a retailer trying to forecast sales for each product.
The events that have the most impact on the retailer are neighborhood events and having a sale in store.
Amazon forcast's Related Time Series Dataset allows us to use promotions per product, such as item_id, timestamp, promotion_applied, but it is difficult to send information for all products in this format.
Amazon forcast has built-in holiday data, but can we use our own store-wide or regional event information in addition to this data?
For example, information such as
2021-8-14, have a summer sale [begin]
2021-8-15, have a summer sale
2021-8-16, have a summer sale
2021-8-16, local festival
2021-8-17, have a summer sale [end]
2021-9-4, local festival
2021-9-24, before school holiday
2021-9-25, school holiday
*Aug-16 has two events on the same day.
If I can have this kind of setup, is it better to specify the beginning and end of the sale? like [begin] or [end].
Generally, sales tend to be larger on the start and end dates of a sale.
Also, if sales are slightly higher on the day before a school holiday, would it be better to include information such as before school holiday?
I have looked into the responses of "ItemSeach ()" and "lookUp()" functions in Amazon Advertising API and
could not find a possible way to get daily/monthly sales of an item.
Popular product research software like , JungleScout, ProfitPhonix, AMZ tracker etc do display Number of monthly sales but all of them show different results.
Does Amazon provide this information ? If not then how the above software are estimating it?
I think when they fetch the ASIN information, they do store "some thing" in their DB and next time when the same ASIN is pulled again then the estimated sales are roughly calculated based on DB previous value/score.
Any help will be highly appreciated .
Thanks
It is not a solution, but here is a reply from UnicornSmasher I found, it may help to save time searching for something that doesn't exist.
constantine We just took all of the bulk data from the products that are being tracked in AMZ Tracker and applied a formula to it all. If you have specific products that are way off please let us know! Certain categories we had less data on. This is version 1 of the research tool, so I'm sure it will continue to improve quickly over time.
Here is the link to question and answer:
amz forum
So, now, the question is 'What formula do they use?'
Let me know if you come up with an idea :)
Let me tell you first that if you're not the part of the Amazon data team you can't get the sales numbers of any product. And, its probably not easy to estimate sales using Amazon advertising API. You need to constantly track a huge number of products to estimate the sales. Here I can explain how AMZ Insight an Amazon tracking tool estimates the sales of any product.
They constantly track a few thousand products from all the categories and collect massive data. Then their in-house data scientist analyze the data to form the sales estimating algorithm. Relationship of multiple data points plots a scattered graph which means of course sales estimates are not 100 percent right.
Data is continuously gathered and analyzed by tracking the Best Seller Rank (BSR), Buybox, reviews and more factors. Then the relationship between this data is formed to come up with the unit sales. Once this relationship is in place then it is much easier to estimate monthly sales and revenue for the product.
I want to create a report in Netsuite ERP that shows me the information about Departments Sales and Budget by Month. I think I can achieve this by creating a saved search that chooses this items, however I don't understand under which category I can find this fields. A saved search would be ideal as I am trying to authomatize the reports in a java application, and I discovered that I can call the savedSearch results.
I found the Department under the standard Criteria in the subcategory "Owner..." and I added a Date standard criteria with the values "within this month", however I have not found the group that contains all the Sales/Income/Margin or the budget (though, I found an aggregation sum function that may be used along with a field). I will appreciate any help. Also, will the addition of this fileds be enough to get the Sales X Department X Date information or do I have to use a different join method?
Thanks!
You'd have to combine two saved searches to achieve this.One on budgets for the period you need. The department column is available on the budgets saved search.
The other would be transactions for the period. Generally budgets are against posting transactions so Invoices, Cash Sales, Credit Memos and Cash Refunds would be in your other search. If you group those by Department you could then combine the two searches in code to create your own budget vs actuals report.