Hello dear Oscar Gonzalez , I hope you have a nice day.

I usually order the matrix with the sum of the rows and columns before entering it in R software, since the default order of the program is to maximize the display of the links between the species, while the other graphical option draw the network as you have ordered from the original matrix.

I hope you find this information useful.

Cristian

Gracias Cristian. You helped a lot. Something else that was needed to be done after ordering the species in the table was to change the default method of cca to normal in the plotweb script. It worked!

Oscar Gonzalez para eso andamos. Jeje pensé que por estar en Florida no hablabas español. Justo es lo que pensé cuando vi tu mensaje que posiblemente tenías el script con la opción de ordenado en cca. Que bien que te sirvió la info.

Saludos,

Cristian

Dear Oscar Gonzalez, you can use sortweb(data, sort.order="dec") function before you using plotweb() function.

e.g.,

data(Safariland)

plotweb(sortweb(Safariland, sort.order="dec"), method="normal") # Decreasing

plotweb(sortweb(Safariland, sort.order="inc"), method="normal") # Increasing

Hai-Dong

Thanks Hai-Dong Li

Hi, Cristian Adrian Martínez-Adriano! How did you turn vertically the graph?

Hi dear Natan Diego Alves de Freitas !!! I hope you have a nice Thursday.

In bipartite package is impossible (at least as far as I know) to modify the network orientation. In this case, what I do is export the image in PDF and I give it the necessary rotation with some editing program. Another option (simpler, I think) is save the image in any png, jpeg, or tiff format and then use the image viewer to rotate them.

BipartiteD3 package is other option that makes interactive interaction networks. I am honest that I have not used it for my graphs, in its options it includes the orientation of the graphic in horizontal and vertical.

Regards,

Cristian

Hi Cristian Adrian Martínez-Adriano thank you, this is helpful and more simple!

Hi everyone! does anyone know how to change the order of the species by their number of links? (e.g. generalists species in the left side and specialist species in the righ side)...

Paulina Corona Tejeda no estoy seguro si entiendo tu pregunta. ¿Quieres colocar las especies con pocos links a la derecha y las de muchos links a la izquierda? Si es así, puedes hacer el arreglo directo en la base de datos. Si usas Excel puedes ordenar las filas y las columnas de menor a mayor según sus totales. De ese modo ya no tienes que modificar nada en bipartite/I'm not sure if I understand your question. Do you want to place the species with few links to the right and those of many links to the left? If so, you can sort the species directly on the database. If you use Excel you can order the rows and columns from highest to lowest basen on their total scores. That way you do not have to modify anything in bipartite package.

Paulina Corona Tejeda Please see my message mentioned above. Plotweb function can be used to sort the species by their number of links.

Como puedo usar el paquete BiMat de matlab que usa la base de bipartite para obtener estos gráficos en R:

Is there any sort of coding to create such a graph? would be very helpful to look at the supplementary documents of the paper. Cheers

Dear Ada Barbanera hope you have a nice day.

I comment to you that Hai-Dong Li gives us (several comments above) an example of the code to sort the interaction network plot.

Regards,

Cristian

Hi Cristian Adrian Martínez-Adriano , I cannot really see any comments about codes. More about specific questions?

Dear Ada Barbanera this is the text shared bye Hai-Dong Li .

Dear Oscar Gonzalez, you can use sortweb(data, sort.order="dec") function before you using plotweb() function.

e.g.,

data(Safariland)

plotweb(sortweb(Safariland, sort.order="dec"), method="normal") # Decreasing

plotweb(sortweb(Safariland, sort.order="inc"), method="normal") # Increasing

Hai-Dong

Another R package, igraph, is much more flexible in terms of drawing options. You may even draw bipartite graphs in it. See the online handbook: https://igraph.org.

Thanks, Marco! I will explore it in my future projects.

You're welcome, Oscar Gonzalez ! You may also check some drawing tutorials focused on this package on my GitHub: https://github.com/marmello77/

Another tip: to ask questions about coding, including R, the best place is StackOverflow, a kind of social network/forum: https://stackoverflow.com. It's really worthy creating an account there.

Yes, Marco. I am familiar with stackoverflow, it is helpful. But the best answers to my questions about R programming were solved by people like you in Research Gate, either by this way or in private messages. Regards.

how do you flip the graph? i.e.the output being vertical (i.e. one group on the left, the other on the right). thanks

Stéphane Cordeau Probably you can use bipartiteD3 package.

Supratim Laha

sort the species first and then use plotweb function.

e.g. higher_seq

That's a helpful recommendation, Hai-Dong. Thanks for sharing.

If you try use bipartieD3 is much better. I recommend you.

Oscar Gonzalez Supratim Laha

Hai-Dong Li when I am using sortweb, I notice that it places the 'species' that contains a missing value last regardless of the overall abundance. Is there a way I can ignore the NA in one column without dropping the entire row? I could specify the order of species precisely using seq but want something I could apply more universally. Many thanks!

K. Ann Bybee-Finley Hi. I do not understand why NAs in your network matrix here. In normally, the cells in matrix should be numbers, such as 0 and 1 in binary network, other numbers for quantitative network . Maybe you can use "0" instead of NA, but not sure. The marginal total is NA if the row of the matrix with NAs due to NA+any numbers = NA.

How can i arrange the data in csv file for plot web? will you give any model?

@Sathish Kumar That depends on your data. but basically it is species (rows) vs interaction (column)

Hello everyone,

I'm currently writing my bachelor thesis which for I need the bipartite package. While I managed to create a few nicely-looking plotwebs - not least thanks to you, Hai-Dong Li – I struggle with turning the horizontal plotwebs upright, as seen in the image attached to the beginning of this thread. Is anyone able to help me with that? Thanks in advance!

You must agregar horiz()

Thank you Asiel Cabrera Guerrero for your immediate answer. Could you please tell me, which package I need to install to run the horiz() function since R does not recognise the command.

Annia Gully-Watson let's use bipartiteD3 :

You need ALL this library:

library(r2d3) library(bipartite ) library(purrr) library(dplyr) library(tidyr) library(stringr) library(tibble) library(RColorBrewer) library(bipartiteD3)

testdata SmallTestWeb

#In data.frame (your data)

bipartite_D3(SmallTestWeb)

bipartite_D3(SmallTestWeb, colouroption = 'brewer', ColourBy = 1, filename = 'demo2')

Put in your R:

??bipartiteD3 and you get all code

library (bipartite)

plotweb(Safariland)

plotweb(Clepto2, method=”normal”, text.rot = 90, labsize =1.5, ybig = 0.7, low.y = 0.7, high.y = 0.98, plot.axes = FALSE, y.width.low = 0.05, y.width.high = 0.05, col.high = “lightgrey”, bor.col.interaction=”black”, bor.col.high=”black”, low.spacing=0.03, high.spacing=0.08)

Here you can adjust at your convenience

Thank you!

I found very useful the code provided by Asiel Cabrera Guerrero. But I am struggling to find out how to change the color of the "col.high = “lightgrey”" and make their interactions change with the same color. Using the code provided they keep as 'light gray' even if I change the color of the column.

Finally managed to get a graph that I can change everything I need...

plotweb(predbi, method="normal", text.rot = 90,

labsize =1.8, ybig = 0.75, low.y = 0.7,

high.y = 1.20, plot.axes = FALSE, y.width.low = 0.05,

y.width.high = 0.05, col.high = "orange",

col.low = "black",

col.interaction = "orange",

bor.low.abun.col = "black",

bor.col.low = "black",

bor.col.interaction="black",

bor.col.high="black",

low.spacing=0.03, high.spacing=0.09)

Hi everybody! Maybe someone knows the correct form for making a graph about antagonism interactions specifically for birds and their hemoparasites. Im looking for information on internet but all the recomendations is for mutualism. Thanks!

Karen Amparo Meneses-Córdoba Maybe you can make a one-part chart. The chord package is good for that.

Asiel Cabrera Guerrero I wrote you to inbox with more details about the data that I want to graph because I dont know "The chord package"... Thanks!