Difference between revisions of "GRA"

From PrimeFish Wiki
Jump to: navigation, search
(Created page with "=WP2: ECONOMIC PERFORMANCE AND PRICES= ---- ---- ==Introduction== ---- :WP2 analyses economic performance of the European fisheries and aquaculture sectors using aggregate da...")
 
 
(12 intermediate revisions by 2 users not shown)
Line 1: Line 1:
=WP2: ECONOMIC PERFORMANCE AND PRICES=
+
 
 +
= Growth Risk Analyser =
 +
 
 
----
 
----
 +
 
----
 
----
==Introduction==
+
 
 +
 
 +
 
 +
== Introduction ==
 +
 
 
----
 
----
  
:WP2 analyses economic performance of the European fisheries and aquaculture sectors using aggregate data obtained from available public sources as well as detailed data from individual companies. This will allow for comparison of the performance of these sectors within Europe, as well as between European countries and other relevant international players. In particular, detailed analysis of growth, productivity and forefront efficiency, using parametric and non-parametric methods, will be conducted on individual European case studies of the chosen species and compared with the performance of Canadian cod and Vietnamese pangasius producers. Additionally, WP2 analyses historically the behaviour of seafood prices in general, as well as the development of market prices of the chosen species, focusing especially on the “boom and bust” cycles characteristics. The outcome of the WP will be critical factors and bottlenecks in the economic performance of the salmon, freshwater trout, cod, sea-bass and bream and herring sectors (SO2).
+
In the research phase of the project, simulation models were developed to analyse how changes in supply and demand affect production planning, economic performance, supply chain relationships, value added, potential product success, market trends and developments and thus competitiveness as measured by the updated [[CPA|FACI]]. One of the outcomes of the research is the Growth risk analyser tool, implemented through simulation/forecasting models for analysing and forecasting pricing trends in the short term: up to 12 months with a confidence band of 95% in the case of this tool. Forecasts given outside of 12 months become less reliable as the forecast horizon increases (up to 24 months are generated by the tool).
 +
 
 +
The tool analyses a time series of data of any value type such as profit per month, average sales quantity per day/week/fortnight/etc, average catch/landings per month/semester/year/etc, containing a minimum of 24 values (to ensure accuracy).
 +
 
 +
 
 +
 
 +
== Tool Overview ==
  
==Objectives==
 
 
----
 
----
  
*Compare the economic performance of selected European pelagic (herring) fisheries.
+
=== Landing page ===
*Compare the economic performance of selected European demersal (cod) fisheries to the performance of the EasternCanadian demersal (cod) fisheries.
+
 
*Compare the economic performance of selected species farmed in Europe (sea bass/sea bream, salmon and freshwatertrout) to pangasius farmed in Vietnam.
+
The tool is very simple and it's homepage presents the users with two different forms of input for the data to be analysed.
*Study the behaviour of seafood prices in general and the development of market prices for the selected species, cod,herring, salmon, trout, sea bass/ sea bream as well as shell fish, focusing especially on the factors characterising the observed “boom and bust” cycles.
+
 
 +
[[File:Grahome.png|GRA home]]
  
==Tasks==
+
The input on the left allows the user to upload a file with the time series to be analysed. The file MUST be on a .CSV format to ensure compatibility with the forecasting algorithm, containing all the values in the first column, one value per row:
----
+
 
 +
<code>value1</code>
 +
 
 +
<code>value2</code>
 +
 
 +
<code>value3</code>
 +
 
 +
<code>value4</code>
 +
 
 +
<code>...</code>
 +
 
 +
Each row represents a time point in the time series. You don't need to enter the month for the values, as each entry is considered a new point value for the month following its preceding value. The results will display the index value of each entry as ordered in the entry data in the X axis starting at 0, and the forecast results will be added as entries at the end of the time series original values.
 +
 
 +
You can create a .CSV file with any spreadsheet editor by simply creating a new spreadsheet and adding the values in the first column, one per row as in the image below:
 +
 
 +
[[File:Sprdsheetvalues.png|time series in a spreadsheet]]
 +
 
 +
After that, click the '''file''' menu of your spreadsheet editor and select '''save as...'''. Once the dialog box for saving opens up, insert the name for the file and select the location. The last and most important and final step is to select '''Text CSV''' as a type in the '''format''' dropdown for the file type which is generally located above the save button. Once you selected the correct type, click the save button and the file will be ready to be uploaded. You can then proceed to the GRA tool home page and click the ''choose file'' button to select your saved file and then click the ''Upload file and calculate forecast'' button to generate your forecast. You can download an example .csv file [[File:Example values.csv|here]].
 +
 
 +
The input on the left allows you to quickly create a forecast by pasting the values of a time series in the text box and using it as an input. Each value for the time series must be entered one after the other, separated by a semi-colon '';''. You will get an error message if any of the values are not in the correct format. Once everything is ready click the calculate button below the input to generate the forecast. You can copy and paste the example below in the tool for a demonstration:
 +
 
 +
<pre>2500;2750;1598;1777;2100;2107;2213;1993;1768;1479;1344;1389;1433;1477;1578;1611;1678;1587;1523;1124;1388;1299;1433;1399</pre>
 +
 
 +
=== The results ===
 +
 
 +
Once you click the forecast button you will be redirected to the results page where you will see on the left sidebar the current request status and the request input as interpreted by the algorithm once the calculations are complete. It may take a few seconds for the algorithm to run, but you should the see the result graphs.
 +
 
 +
[[File:Graresult1.png|GRA results page]] [[File:Graresult2.png|GRA results page]] [[File:Graresult3.png|GRA results page]]
 +
 
 +
The first graph shows the input values and a fitted trend line for the approximation trend generated. The second graph shows the time series residuals, representing the confidence of the forecast. The more values between the confidence bands, the more accurate will be the results. The last graph will show the original input, the forecast trend line with the forecast values for the next 24 months and the confidence bands. Remember the algorithm is only effective for the first 12 months in its projected confidence, and the further the value goes in the prediction after the 12th value the less accurate is the prediction and should not be considered as having the same forecast confidence.
 +
 
 +
&nbsp;
  
#Economic performance of selected individual sector.
+
== References & Readings ==
**Demersal (cod) fisheries
 
**North Atlantic pelagics (herring)
 
**Freshwater trout
 
**Atlantic salmon
 
**European seafood market
 
#Identifying and characterising “boom and bust” cycles.
 
**Boom and bust” cycles.
 
**Impact of macro-economic effects on “boom-and-bust” cycles
 
**Price transmission and market integration
 
  
==Deliverables==
 
 
----
 
----
 
*Report on the development of prices & volumes in the European fishery & aquaculture market
 
*Report on the economic performance of selected European and Canadian fisheries
 
*Report on the economic performance of selected European and Vietnamese farmed species
 
*Report on “boom and bust” cycles for selected European fisheries and aquaculture species
 
*Manuscript to a peer-reviewed journal on “boomand-bust” cycles in European seafood markets
 

Latest revision as of 18:03, 5 August 2018

Growth Risk Analyser



 

Introduction


In the research phase of the project, simulation models were developed to analyse how changes in supply and demand affect production planning, economic performance, supply chain relationships, value added, potential product success, market trends and developments and thus competitiveness as measured by the updated FACI. One of the outcomes of the research is the Growth risk analyser tool, implemented through simulation/forecasting models for analysing and forecasting pricing trends in the short term: up to 12 months with a confidence band of 95% in the case of this tool. Forecasts given outside of 12 months become less reliable as the forecast horizon increases (up to 24 months are generated by the tool).

The tool analyses a time series of data of any value type such as profit per month, average sales quantity per day/week/fortnight/etc, average catch/landings per month/semester/year/etc, containing a minimum of 24 values (to ensure accuracy).

 

Tool Overview


Landing page

The tool is very simple and it's homepage presents the users with two different forms of input for the data to be analysed.

GRA home

The input on the left allows the user to upload a file with the time series to be analysed. The file MUST be on a .CSV format to ensure compatibility with the forecasting algorithm, containing all the values in the first column, one value per row:

value1

value2

value3

value4

...

Each row represents a time point in the time series. You don't need to enter the month for the values, as each entry is considered a new point value for the month following its preceding value. The results will display the index value of each entry as ordered in the entry data in the X axis starting at 0, and the forecast results will be added as entries at the end of the time series original values.

You can create a .CSV file with any spreadsheet editor by simply creating a new spreadsheet and adding the values in the first column, one per row as in the image below:

time series in a spreadsheet

After that, click the file menu of your spreadsheet editor and select save as.... Once the dialog box for saving opens up, insert the name for the file and select the location. The last and most important and final step is to select Text CSV as a type in the format dropdown for the file type which is generally located above the save button. Once you selected the correct type, click the save button and the file will be ready to be uploaded. You can then proceed to the GRA tool home page and click the choose file button to select your saved file and then click the Upload file and calculate forecast button to generate your forecast. You can download an example .csv file File:Example values.csv.

The input on the left allows you to quickly create a forecast by pasting the values of a time series in the text box and using it as an input. Each value for the time series must be entered one after the other, separated by a semi-colon ;. You will get an error message if any of the values are not in the correct format. Once everything is ready click the calculate button below the input to generate the forecast. You can copy and paste the example below in the tool for a demonstration:

2500;2750;1598;1777;2100;2107;2213;1993;1768;1479;1344;1389;1433;1477;1578;1611;1678;1587;1523;1124;1388;1299;1433;1399

The results

Once you click the forecast button you will be redirected to the results page where you will see on the left sidebar the current request status and the request input as interpreted by the algorithm once the calculations are complete. It may take a few seconds for the algorithm to run, but you should the see the result graphs.

GRA results page GRA results page GRA results page

The first graph shows the input values and a fitted trend line for the approximation trend generated. The second graph shows the time series residuals, representing the confidence of the forecast. The more values between the confidence bands, the more accurate will be the results. The last graph will show the original input, the forecast trend line with the forecast values for the next 24 months and the confidence bands. Remember the algorithm is only effective for the first 12 months in its projected confidence, and the further the value goes in the prediction after the 12th value the less accurate is the prediction and should not be considered as having the same forecast confidence.

 

References & Readings