Page 444 - Contributed Paper Session (CPS) - Volume 4
P. 444
CPS2564 Tiffany Rizkika et al.
survey is done every week for volatile commodities, and every two weeks for
non-volatile products. The publication of this data is published every month.
In making decisions related to fluctuating food prices, data with a
collection time very close to the time of decision making is needed. The closer
the time span, the better decisions can be made. Currently the consumer
prices are published every month. Nowadays when quick decision making on
prices with rapid changes, more frequent and almost real time data and
prediction so called now casting is needed. The term "nowcasting" is originally
used in meteorology to forecast weather in the present and in the next few
hours. Several approaches can be implemented to monitor the food prices in
almost real-time to maintain price transparency therefore government will be
easy detecting inflation early to reach the stability of economy.
Pramana, et al. (2016) shows a proof of concept of using crowdsourcing
data as one resource for food price nowcasting. They also discuss that in
general there is a similar movement of official price data patterns between
several commodities that collected by BPS-Statistics Indonesia, with data
collected through crowdsourcing techniques carried out by Pulse Lab Jakarta.
They used simple data cleaning and smoothing for nowcasting and only few
commodities.
One of big challenges in big data is its veracity and followed by the
analysis. Hence, further research on nowcasting technique required, including
filtering and modelling, to obtain accurate prediction is needed. Therefore, in
this study, several nowcasting approaches are investigated to obtain the best
nowcast for food price data.
2. Methodology
a. Data Sources. This study uses the following secondary sources:
1) Crowdsourcing data obtained from collaboration of Pulse Lab
Jakarta, United Nation’s Food and Organization (FAO), and
Premise Data Corporation in 2015.
2) Market data obtained from google to identify places/markets
3) Official Consumer Price Data, from BPS Statistics Indonesia
consumer price survey.
b. Food Commodities:
1) Volatile: chicken beef, egg, onion, chili, low quality rice, and
premium quality rice.
2) Non-volatile: mackerel, long bean, instant dry noodles,
peanuts, and vegetable tomatoes.
c. This study is taking place at Lombok Island, during March until July
2015.
433 | I S I W S C 2 0 1 9