Finding slope for the polyfit line

load data_matrix.mat

data=SECTION_L;

data(:,1:3)=[];

row=9;

cleanup=data([1 2 3 row],:);

% year range

y1 = 1975;

y2 = 2016;

% init

x = y1:y2;

nx = numel(x);

mo_av = zeros(nx,12);

mo_max = zeros(nx,12);

mo_min = zeros(nx,12);

max_day = zeros(nx,12);

min_day = zeros(nx,12);

k = 0;

for k = 1:nx % year loop

col_dat_yr=find(cleanup(1,:)== x(k)); % find data corresponding to year = x(k)

data_yr=cleanup(:,col_dat_yr);

for mmonth = 1:12

col_dat_mo=find(data_yr(2,:)==mmonth);

data_mo=data_yr(:,col_dat_mo);

mo_av(k,mmonth)=mean(data_mo(4,:));

mo_max(k,mmonth)=max(data_mo(4,:));

mo_min(k,mmonth)=min(data_mo(4,:));

max_day(k,mmonth)=find(data_mo(4,:)==mo_max(k,mmonth));

min_day(k,mmonth)=find(data_mo(4,:)==mo_min(k,mmonth));

end

end

% plot montly data

str_mo = {'Jan' 'Feb' 'Mar' 'Apr' 'May' 'Jun' 'Jul' 'Aug' 'Sep' 'Oct' 'Nov' 'Dec'};

figure

sgtitle('Monthly Data : Avg')

for ck = 1:12

coefs1=polyfit(x,mo_av(:,ck),1);

curve1=polyval(coefs1,x);

subplot(3,4,ck)

plot(x,mo_av(:,ck),'-',x,curve1);

xlim([1970,2020])

ylim([-25,5])

xlabel('Years')

ylabel('Temperature C')

title(str_mo(ck));

end

figure

sgtitle('Monthly Data : Max')

for ck = 1:12

coefs2=polyfit(x,mo_max(:,ck),1);

curve2=polyval(coefs2,x);

subplot(3,4,ck)

plot(x,mo_max(:,ck),'-',x,curve2);

xlim([1970,2020])

ylim([0,25])

xlabel('Years')

ylabel('Temperature C')

title(str_mo(ck));

end

figure

sgtitle('Monthly Data : Min')

for ck = 1:12

coefs3=polyfit(x,mo_min(:,ck),1);

curve3=polyval(coefs3,x);

subplot(3,4,ck)

plot(x,mo_min(:,ck),'-',x,curve3);

xlim([1970,2020])

ylim([-30,10])

xlabel('Years')

ylabel('Temperature C')

title(str_mo(ck));

end

I have made the plots with 12 graphs for three figures and have put a line through the 36 graphs. How would you find the slope of the polyfit and polyval lines for every graph.

Answer 1

[Voss]

polyfit returns the coefficients in a vector, with the highest-degree coefficient first. Therefore, the slopes are the values of coefs1(1) (similarly, coefs2(1) and coefs3(1)) in each iteration.

[Michelangelo Cannistraro]

but i need to get 36 slopes and my variable curve1 etc only give one slope

[Torsten]

Replace

coefs1=polyfit(x,mo_av(:,ck),1);

by

coefs1=polyfit(x,mo_av(:,ck),1);
slope1(ck) = coefs1(1);

(same for the remaining two loops).

Then you have 36 slopes (slope1(1:12),slope2(1:12) and slope3(1:12)).