forecasting cement prices
TRANSCRIPT
-
8/13/2019 Forecasting Cement Prices
1/14
Team 4
Amit Tyagi
Harmanjit Singh
Mohini Jain
Nandini ChandrasekharRahul Chakraborty
-
8/13/2019 Forecasting Cement Prices
2/14
Objective
Data
Background
Quick Look
Issues
Preparation
Analysis
Naive
Exponential Moving Average
Time Series Forecasting
OBJECTIVE
DATA
RESULTS SUMMARY & INSIGHTS
ANALYSIS
-
8/13/2019 Forecasting Cement Prices
3/14
Industry Very cost intensive Major raw materials used for construction include sand, cement, steel etc. Prices of raw materials are driven by the price of crude oil, indicating
fluctuation in price of the raw materials.
Purpose
To predict cement prices on monthly basis in context of India.
Stakeholders
Builders/dealers in the construction industry or cement productioncompanies to plan capacity expansion
Benefits
The forecasts will serve as a valuable tool to help the construction industrymake important logistic decisions such as:
Timing of purchase Inventory Management
Strategically allocate budget inpurchasing cement
In effect, the outcome would be better resource planning and sourcingleading to increased savings
The outcome would be better resource planning and sourcing leading to increased savings
INDUSTRY
PURPOSE
STAKEHOLDERS
BENEFITS
-
8/13/2019 Forecasting Cement Prices
4/14
Sourc e: Monthly Crude oil and cement prices from www.indiastat.com
Time Period: Data available for the period Jan 2009 May 2011
Forecast Peri June 2011
Aug 2011
Based on monthly prices of cement and oil for the years 2009 and 2010
Average Price of Cement in Major Consumption (Centers Per Bag of 50 kg (In Rs.241)
Monthly Average Price of Indian Basket of Crude Oil (in $/barrel converted to Rs3471)
Crude oil to be explored as a predictor for cement prices
SOURCE
TIME PERIOD
FORECAST
http://www.indiastat.com/http://www.indiastat.com/ -
8/13/2019 Forecasting Cement Prices
5/14
Gaps/Issues in data Cement prices were missing for the data period Jan 2011 to Mar 2011
Oil Prices were available in US dollar Correlation between Cement Prices and Oil Prices was low at 0.33
Moving Average
Missing data imputation in cement prices using the method of MovingAverage and straight line method
Oil prices converted to Indian Rupees using historic conversion ratecorresponding to each data point sourced from www.oanda.com
http://www.oanda.com/http://www.oanda.com/ -
8/13/2019 Forecasting Cement Prices
6/14
-
8/13/2019 Forecasting Cement Prices
7/14
Cement
Level: Averages around 237.5
Trend:No clear Trend present in the series
Seasonality: No significant pattern
Stationary series
Peak around the month of April every year
Oil
Level:for oil prices is around 3500
Trend: Clear increasing trend
Seasonality: No significant pattern
Non Stationary Series
2010 values follow a very smooth pattern
-
8/13/2019 Forecasting Cement Prices
8/14
200.00
210.00
220.00
230.00
240.00
250.00
260.00
270.00
280.00
290.00
Cement Prices Nave Forecast
Parameter Value
Average Error -0.30
MAE 12.68
RMSE 16.04
MAPE -0.43%
Time
Forecasted
Cement Prices
June. 2011 240.00
July. 2011 233.00
Aug. 2011 225.00
Cement data displays monthly seasonalityNave 12-month ahead forecastForecast values and Error Measurements attached below
ValidationPeriod
-
8/13/2019 Forecasting Cement Prices
9/14
Cement data does not depict trend but has 12 month seasonality Holt-Winter No Trend method of Exponential Smoothing
Data partitioned into 24 months Training and 5 months Validation
Forecasts on validation set:
MAPE -0.51%
MAE 19.24
RMSE 10.78
Time Actual Forecast Error LCI UCI
Jan. 2011 237.67 240.10 -2.43 219.09 261.11
Feb. 2011 238.56 237.66 0.90 216.65 258.67
Mar. 2011 238.69 233.27 5.42 212.26 254.27
Apr. 2011 280.00 231.33 48.67 210.33 252.34
May. 2011 276.00 237.24 38.76 216.23 258.24
Time Forecast LCI UCI
Jun. 2011 240.10 219.09 261.11
July. 2011 237.66 216.65 258.67
Aug. 2011 233.27 212.26 254.27
The RMSE error is lower for the
method than Nave forecasting.
Both Exponential and Navemethod systematically under
predicts the cement demand
200
210
220
230
240
250
260
270280
290
CementPrices
Time
Time Plot of Actual Vs Forecast (Training Data)
Actual Forecast
ValidationPeriod
-
8/13/2019 Forecasting Cement Prices
10/14
Independent Variables : time & Dummy variables for Q1, Q2 and Q3 ( Q4 as Base Quarter)
200210220230240250260
Jan.2009
Mar.2009
May.2009
July.2009
Sept.2009
Nov.2009
Jan.2010
Mar.2010
May.2010
July.2010
Sept.2010
Nov.2010
Predicted Value Actual Value
210
230
250
270
290
Jan. 2011 Feb. 2011 Mar. 2011 Apr. 2011 May. 2011
The RMSE error is lower for the method
than previous methods.
However the method systematically
under predicts in validation set
Adjusted R 2 for the model is .3345
TestData
ValidationData
-
8/13/2019 Forecasting Cement Prices
11/14
Independent Variables : time & Dummy variables for each Month( Dec. as Base Month)
The RMSE error is lower for the method
than previous methods.
Method has similar errors in validation
and test sets
Adjusted R 2 for the model is .5364
TestData
ValidationData
200
210
220
230
240
250
260
Predicted Value Actual Value
220
240
260280
300
Jan. 2011 Feb. 2011Mar. 2011 Apr. 2011 May.
2011
-
8/13/2019 Forecasting Cement Prices
12/14
Independent Variables : Change in Crude oil & Dummy variables for each Month( Dec. asBase Month)
The RMSE error is lower for the method
than previous methods.
Method systematically under predicts in
validation/Test Set
Adjusted R 2 for the model is .5964
TestData
ValidationData
200
220
240
260
Predicted Value Actual Value
220
230
240
250
260
270
280
290
Jan. 2011 Feb. 2011 Mar. 2011 Apr. 2011 May. 2011
-
8/13/2019 Forecasting Cement Prices
13/14
Based on Model 3 the forecast for June 2011 is:
Point Estimate: 244.67
95% confidence interval is (230.50,258.9)
-1
-0.5
0
0.5
1
1 2 3 4 5 6
Number of Lags
Autocorrelation of Residual / Data
Set #1
-
8/13/2019 Forecasting Cement Prices
14/14
April and May have at least 10%
higher cement prices than December
Cement Prices are affected byprevious months change in Oil pricesand not oil prices themselves
Crude prices have a positive trendoverall last two years.
Inventory should be brought inDecember for next fiscal yearrather than April/May
Track changes in crude prices onmonthly basis to deduce future
cement prices
INSIGHTS RECOMMENDATIONS