adding to the technical deployment guide
This commit is contained in:
Tomas Singliar
Родитель
5dd9fba794
Коммит
a985cd21c7
|
|
@ -37,7 +37,7 @@ This solution analyzes your prices and makes recommendations,
|
|||
|
||||
All information is provided the fine level at which you need to control your price and inventory.
|
||||
|
||||
More detailed description is available in our
|
||||
Additional detail on the data science of prices are in our
|
||||
[blog post](https://blogs.msdn.microsoft.com/intel/archives/1015).
|
||||
|
||||
# Solution Architecture
|
||||
|
|
|
|||
|
|
@ -282,19 +282,19 @@ CREATE TABLE dbo.Forecasts (
|
|||
ChannelName varchar(100) not null,
|
||||
CustomerSegment varchar(100) not null,
|
||||
LastDayOfData date not null, -- forecast made using data up and including this
|
||||
PeriodInDays int not null,
|
||||
PeriodsAhead int not null,
|
||||
ForecastPeriodStart date not null,
|
||||
ForecastPeriodEnd date not null, -- end of the period whose demand is forecasted
|
||||
UnitPrice float, -- forecast is conditional on this price. Should be decimal (6,2), but that's pulling the ADF tiger's tail.
|
||||
PeriodInDays int not null,
|
||||
PeriodsAhead int not null,
|
||||
ForecastPeriodStart date not null,
|
||||
ForecastPeriodEnd date not null, -- end of the period whose demand is forecasted
|
||||
UnitPrice float, -- forecast is conditional on this price.
|
||||
Demand float not null,
|
||||
Demand90LB float not null,
|
||||
Demand90UB float not null,
|
||||
ActualSales float null,
|
||||
sAPE float null,
|
||||
qBar float null
|
||||
ActualSales float null,
|
||||
sAPE float null,
|
||||
qBar float null
|
||||
primary key (RunDate, Item, SiteName, ChannelName,
|
||||
LastDayOfData, ForecastPeriodStart, ForecastPeriodEnd)
|
||||
LastDayOfData, ForecastPeriodStart, ForecastPeriodEnd)
|
||||
)
|
||||
```
|
||||
|
||||
|
|
@ -310,10 +310,7 @@ To put the sAPE in perspective, the qBar is a smoothed measure demand around the
|
|||
#### Suggestions dataset
|
||||
|
||||
The SuggestionRuns table stores the pricing suggestions made from
|
||||
the elasticities and forecasts. The suggestionRunID is an identifier
|
||||
referring to the date of model build from which the suggestion are created.
|
||||
|
||||
The
|
||||
the elasticities and forecasts.
|
||||
|
||||
```sql
|
||||
CREATE TABLE [dbo].[SuggestionRuns] (
|
||||
|
|
@ -345,27 +342,70 @@ CREATE TABLE [dbo].[SuggestionRuns] (
|
|||
);
|
||||
```
|
||||
|
||||
<div class="todo" style="color:red; font-weight: bold;">
|
||||
TODO: Describe the table columns
|
||||
</div>
|
||||
The suggestionRunID is an identifier referring to the date of model
|
||||
build from which the suggestion are created.
|
||||
|
||||
<tt>PastPeriodStart</tt> and <tt>PastPeriodEnd</tt> describe the time interval for which
|
||||
the baseline numbers (Orders, Revenue, Margin) are taken (the "past period").
|
||||
<tt>SuggestionPeriodStar</tt> and <tt>SuggestionPeriodEnd</tt> describe the period for
|
||||
which the price is proposed (the "suggestion period").
|
||||
<tt>minOrders</tt> is the minimum number of orders that need to have occured
|
||||
for the item in the past period to be considered in the suggestion pipeline.
|
||||
|
||||
Then we have the baseline numbers: <tt>UnitsLastPeriod</tt>, <tt>avgSaleUnitPrice</tt>,
|
||||
<tt>avgCostUnitPrice</tt>, <tt>RevenueLastPeriod</tt>, <tt>MarginLastPeriod</tt>
|
||||
whose interpretations are hopefully clear.
|
||||
<tt>Orders</tt> will be more than 1 only if disaggregated data are entered.
|
||||
|
||||
<tt>Elasticity</tt> comes from the model estimatuon step, and the optimal prices
|
||||
follow from it and the marginal cost (avgCostUnitPrice). The exact price maximizing
|
||||
the gross profit margin is <tt>marginOptimalPrice</tt>, which is then rounded to
|
||||
"x.y9". The model predicts that it should be possible to make additional
|
||||
<tt>incrementalMargin</tt> dollars over the suggestion period.
|
||||
|
||||
## Configuration
|
||||
|
||||
To set the parameters, update the table <tt>dbo.Parameters</tt> in the Solution's
|
||||
SQL database. These parameters are configurable:
|
||||
SQL database, which stores simple key-value pairs:
|
||||
|
||||
* Lead-time (days before start of pricing period) for producing the pricing suggestions.
|
||||
Insert the key-value pair ('sugLeadTime', '[n]') into the table, replacing [n]
|
||||
by the number of days in advance. The default value is '1'.
|
||||
* Maximum allowed deviation of suggested price from current price (percentage).
|
||||
Please insert the key-value pair ('maxPriceDeviation', '[x.x]') into the table,
|
||||
replacing [x.x] by the desired fraction. The default value is '0.2', representing
|
||||
a maximum deviation of 20 percent.
|
||||
```sql
|
||||
CREATE TABLE [dbo].[Parameters] (
|
||||
[paramName] VARCHAR(50) NOT NULL,
|
||||
[paramValue] VARCHAR(MAX) NULL,
|
||||
PRIMARY KEY ([paramName])
|
||||
)
|
||||
```
|
||||
|
||||
<div class="todo" style="color:red; font-weight: bold;">
|
||||
TODO: describe the parameters table and check parameter names
|
||||
</div>
|
||||
The bulk services are the only ML Services running in the ADF
|
||||
pipeline and therefore need configuration.
|
||||
|
||||
### Recognized parameters for bulk services
|
||||
|
||||
|Parameter (paramName) | Meaning | Default paramValue |
|
||||
|--|--|--|
|
||||
|BulkElasticities_DeltaX | Elasticity for what change in price*? | -0.1 |
|
||||
|BulkElasticities_WeekJump | Retrieve elasticity for every n-th week | 1 |
|
||||
|BulkForecasts_periodsAhead | Forecasts for many periods ahead? | 1 |
|
||||
|BulkCrossPrice_WeekJump |Retrieve elasticity for every n-th week | 1 |
|
||||
|
||||
### Adjusting suggestion lead times
|
||||
|
||||
Today, you can adjust the lead times on suggestion by manipulating the date parameters
|
||||
which Azure Data Factory passes to the <tt>spRecommendProducts</tt> stored procedure Activity
|
||||
(see ADF description below). If you would like multiple suggestion periods, please
|
||||
duplicate the stored procedure activity and call it with different parameters.
|
||||
|
||||
```json
|
||||
"typeProperties": {
|
||||
"storedProcedureName": "spRecommendProducts",
|
||||
"storedProcedureParameters": {
|
||||
"SliceEnd": "$$Text.Format('{0:yyyy-MM-dd}', SliceEnd)",
|
||||
"lastDayOfData": "$$Text.Format('{0:yyyy-MM-dd}', SliceEnd)",
|
||||
"suggestionPeriodStart": "$$Text.Format('{0:yyyy-MM-dd}', Date.AddDays(SliceEnd,1))",
|
||||
"suggestionPeriodEnd": "$$Text.Format('{0:yyyy-MM-dd}', Date.AddDays(SliceEnd,7))",
|
||||
"minOrders": "1"
|
||||
}
|
||||
```
|
||||
|
||||
## Building Applications
|
||||
|
||||
|
|
@ -420,13 +460,16 @@ The same information is reflected in the VIEW SCHEMA pane of the AzureML plugin.
|
|||
There are three types of ML service in this solution, batch model build, interactive retrieval
|
||||
and bulk retrieval services.
|
||||
|
||||
The batch model build service is BuildModel and is responsible for all estimation tasks.
|
||||
Depending on data size, it can run minutes to hours.
|
||||
The batch model build service is BuildModel and is responsible for all estimation
|
||||
and forecasting tasks. Depending on data size, it can run several minutes to hours.
|
||||
|
||||
The interactive services are:
|
||||
* Elasticities
|
||||
* CrossElasticities
|
||||
* Forecasts
|
||||
* Elasticities - retrieve elasticities for one product at all sites, channels, and segments
|
||||
* CrossElasticities - retrieve cross-elasticities for all products and channels at one site.
|
||||
The model assumes the same items at different sites don't compete.
|
||||
Perhaps more questionably, it also assumes that customer segmentation
|
||||
boundaries are not permeable.
|
||||
* Forecasts - retrieve forecasts at one site, assa specific pricing point
|
||||
* PromoSimulation
|
||||
* Outliers
|
||||
* RetrospectiveAnalysis
|
||||
|
|
@ -439,14 +482,30 @@ The bulk services are used to export the data from the model to the database.
|
|||
* BulkCrossElasticities
|
||||
* BulkForecasts
|
||||
|
||||
### Storage and ADF architecture
|
||||
### Storage and ADF structure
|
||||
|
||||
The Azure Data Factory has three Pipelines:
|
||||
- Configure Services Pipeline. This creates small datasets containing solution parameters.
|
||||
- Pricing This is the large pipeline which
|
||||
- * Prepares the data for modeling
|
||||
* Runs the models
|
||||
* Extracts forecasts and elasticities from the model
|
||||
* Loads the model outputs into the database for visualization
|
||||
- Suggestions Pipeline creates pricing suggestions based on outputs of the Pricing pipeline.
|
||||
|
||||
<div class="todo" style="color:red; font-weight: bold;">
|
||||
TODO: describe the parameters table and check parameter names
|
||||
TODO: describe activities in the pipelines
|
||||
</div>
|
||||
|
||||
- ADF pipeline
|
||||
- Datasets
|
||||
The storage account has the following important folders:
|
||||
- crosselasticity - extracted from the model in Pricing Pipeline
|
||||
- elasticity - extracted from the model in Pricing Pipeline
|
||||
- experimentoutput - AzureML experiment internal cache
|
||||
- forecasts - extracted from the model in Pricing Pipeline
|
||||
- originaldata - cleaned, processed input data at start of Pricing Pipeline
|
||||
- pricing - contains the log files from model runs
|
||||
- serviceparameters - parameter datasets produced by the Configure Services pipeline
|
||||
|
||||
|
||||
|
||||
## Troubleshooting
|
||||
|
|
@ -459,7 +518,7 @@ need to use higher tiers of the resources.
|
|||
The blob name is crezted from the datasetName given in the spreadsheet.
|
||||
Open the storage account and make sure a model with the given datasetName exists.
|
||||
The name is case-sensitive.
|
||||
The default datasetName used by ADF is latestModelBuild
|
||||
The default datasetName used by ADF is "latestDemoBuild".
|
||||
|
||||
#### Timeouts on the BuildModel service
|
||||
Try increasing the timeout period in the <tt>retrain_AzureML_Model</tt> ADF activity.
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче